DATA FORWARDING METHOD, DEVICE, AND SYSTEM IN SOFTWARE-DEFINED NETWORKING
20170310588 · 2017-10-26
Assignee
Inventors
Cpc classification
H04L12/1854
ELECTRICITY
H04L45/50
ELECTRICITY
International classification
Abstract
Data forwarding method, device, and system are provided for software-defined networking. An SDN control device delivers a proactive flow entry according to network topology information, where the proactive flow entry is unrelated to a service, and instead related to the network topology information. The proactive flow entry may be reused during data forwarding. After data is received, the SDN control device delivers reactive flow entries to some switching devices, where the reactive flow entries are related to the service, but are delivered to a subset of switching devices. This reduces flow entries delivered by the SDN control device to some switching devices, so that occupation of resources of the SDN control device and a switching device by flow entries is reduced.
Claims
1. A software-defined networking (SDN) system for implementing data forwarding, comprising: an SDN control device; and n switching devices, a first of the n switching devices being connected to a source end of data, an n.sup.th switching device being connected to a destination end of the data, and n≧3, wherein: the first switching device is configured to request a forwarding rule from the SDN control device, the request including characteristic information of the data, and to receive a first forwarding rule returned by the SDN control device, the first forwarding rule including information about a path from the first switching device to the n.sup.th switching device, the first switching device being configured to send to a second switching device connected to the first switching device, the data and the information about the path from the first switching device to the n.sup.th switching device according to the first forwarding rule; and the second switching device is configured to receive the data and the information about the path from the first switching device to the n.sup.th switching device sent by the first switching device, query according to a signpost of a third switching device in the information about the path from the first switching device to the n.sup.th switching device, a second forwarding rule stored in the second switching device, and obtain a port of a third switching device on the second switching device and forward the data to the third switching device through the port on the second switching device.
2. The SDN system according to claim 1, wherein: the information about the path from the first switching device to the n.sup.th switching device comprises signposts of the third switching device to the n.sup.th switching device that are on a data forwarding path; and a match field of the second forwarding rule is the signpost of the third switching device.
3. The SDN system according to claim 2, wherein when n≧4, the second switching device is further configured to send, to the third switching device, information about a path from the second switching device to the n.sup.th switching device, the information about the path from the second switching device to the n.sup.th switching device includes signposts of a fourth switching device to the n.sup.th switching device that are on the data forwarding path.
4. The SDN system according to claim 3, wherein an i.sup.th switching device is configured to receive the data and information about a path from an i−1.sup.th switching device to the n.sup.th switching device that are sent by the i−1.sup.th switching device, wherein the information about the path from the i−1.sup.th switching device to the nth switching device comprises signposts of an i+1.sup.th switching device to the n.sup.th switching device that are on the data forwarding path, i being an integer greater than or equal to 3 and less than n; and the i.sup.th switching device is configured to query, according to a signpost of the i+1.sup.th switching device, an i.sup.th forwarding rule stored in the i.sup.th switching device, and according to the i.sup.th forwarding rule, determine a port of the i+1.sup.th switching device on the i.sup.th switching device and forward the data to the i+1.sup.th switching device through the port on the i.sup.th switching device.
5. The SDN system according to claim 4, wherein if the i+1.sup.th switching device is located between the first and n.sup.th switching device, the i.sup.th switching device being configured to send, to the i+1.sup.th switching device, information about a path from the i.sup.th switching device to the n.sup.th switching device according to the i.sup.th forwarding rule, wherein the information about the path from the i.sup.th switching device to the n.sup.th switching device comprises signposts of an i+2.sup.th switching device to the n.sup.th switching device that are on the data forwarding path, the i+2.sup.th switching device being connected to the i+1.sup.th switching device.
6. The SDN system according to claim 5, wherein the i.sup.th switching device is configured to send to the i+1.sup.th switching device information about a path from the i.sup.th switching device to the n.sup.th switching device according to the i.sup.th forwarding rule; and the i.sup.th switching device is configured to remove the signpost of the i+1.sup.th switching device from the information about the path from the i−1.sup.th switching device to the n.sup.th switching device to obtain the signposts of the i+2.sup.th switching device to the n.sup.th switching device, and send, to the i+1.sup.th switching device, the signposts of the i+2.sup.th switching device to the n.sup.th switching device, according to the i.sup.th forwarding rule.
7. The SDN system according to claim 2, wherein: the first switching device is configured to successively push the signposts of the n.sup.th switching device to the third switching device that are on the data forwarding path into a protocol stack, wherein a signpost of the n.sup.th switching device is shifted into the bottom of the stack, and the signpost of the third switching device is shifted into the top of the stack, so that the protocol stack carries the information about the path from the first switching device to the n.sup.th switching device.
8. The SDN system according to claim 7, wherein the protocol stack comprises a Multiprotocol Label Switching (MPLS) stack, and one MPLS header of the MPLS stack carries a signpost of one of the n switching devices.
9. The SDN system according to claim 1, wherein: the second switching device is configured to identify, according to the signpost of the third switching device, that a next-hop switching device of the data is the third switching device, wherein the signpost of the third switching device comprises at least one of the following: a device identifier of the third switching device, an identifier of the port on the second switching device, and an identifier of a link between the third switching device and the second switching device.
10. The SDN system according to claim 1, wherein: the first switching device is further configured to send, according to the first forwarding rule, a tenant identifier to the second switching device connected to the first switching device, so as to send the tenant identifier to the n.sup.th switching device by using the second switching device; and the n.sup.th switching device is configured to determine the destination end of the data according to the tenant identifier and the characteristic information of the data, and send the data to the destination end.
11. A method for implementing data forwarding, wherein the method is applied to a software-defined networking (SDN) system, the SDN system comprising an SDN control device and n switching devices, a first switching device being connected to a source end of data, an n.sup.th switching device is connected to a destination end of the data, and n≦3, wherein the method comprises: requesting, by the first switching device, a forwarding rule from the SDN control device, wherein the request carries characteristic information of the data; receiving, by the first switching device, a first forwarding rule returned by the SDN control device, wherein the first forwarding rule comprises information about a path from the first switching device to the n.sup.th switching device; sending, by the first switching device, to a second switching device connected to the first switching device, the data and the information about the path from the first switching device to the n.sup.th switching device according to the first forwarding rule; receiving, by the second switching device, the data and the information about the path from the first switching device to the n.sup.th switching device, querying, according to a signpost of a third switching device in the information about the path from the first switching device to the n.sup.th switching device, a second forwarding rule stored in the second switching device, and obtaining a port, on the second switching device, of the third switching device and forwarding the data to the third switching device through the port on the second switching device.
12. The method according to claim 11, wherein: the information about the path from the first switching device to the n.sup.th switching device comprises signposts of the third switching device to the n.sup.th switching device that are on a data forwarding path; and a match field of the second forwarding rule is the signpost of the third switching device.
13. The method according to claim 12, wherein when n≧4, the second switching device sends, to the third switching device, information about a path from the second switching device to the n.sup.th switching device, the information about the path from the second switching device to the n.sup.th switching device comprising signposts of a fourth switching device to the n.sup.th switching device on the data forwarding path.
14. The method according to claim 13, wherein an i.sup.th switching device receives data and information about a path from an i−1.sup.th switching device to the n.sup.th switching device that are sent by the i−1.sup.th switching device, wherein the information about the path from the i−1.sup.th switching device to the n.sup.th switching device comprises signposts of an i+1.sup.th switching device to the n.sup.th switching device on the data forwarding path, i being an integer greater than or equal to 3 and less than n; and the i.sup.th switching device queries, according to a signpost of the i+1.sup.th switching device, an i.sup.th forwarding rule stored in the i.sup.th switching device, and according to the i.sup.th forwarding rule, determines a port, on the i.sup.th switching device, of the i+1.sup.th switching device, and forwards the data to the i+1.sup.th switching device through the port, on the i.sup.th switching device, of the i+1.sup.th switching device.
15. The method according to claim 14, wherein if the i+1.sup.th switching device is located between the first switching device and the n.sup.th switching device, the i.sup.th switching device sending, to the i+1.sup.th switching device, information about a path from the i.sup.th switching device to the n.sup.th switching device according to the i.sup.th forwarding rule, wherein the information about the path from the i.sup.th switching device to the n.sup.th switching device comprises signposts of an i+2.sup.th switching device to the n.sup.th switching device that are on the data forwarding path.
16. The method according to claim 15, wherein sending by the i.sup.th switching device to the i+1.sup.th switching device, information about a path from the i.sup.th switching device to the n.sup.th switching device according to the i.sup.th forwarding rule, comprises: removing, by the i.sup.th switching device, the signpost of the i+1.sup.th switching device from the information about the path from the i−.sup.th switching device to the n.sup.th switching device to obtain signposts of the i+2.sup.th switching device to the n.sup.th switching device, and sending to the i+1.sup.th switching device, the signposts of the i+2.sup.th switching device to the n.sup.th switching device, according to the i.sup.th forwarding rule.
17. The method according to claim 12, wherein: the first switching device successively pushes the signposts of the n.sup.th switching device that are on the data forwarding path into a protocol stack to the third switching device, wherein a signpost of the n.sup.th switching device is shifted to a lower portion of the stack, and the signpost of the third switching device is shifted to an upper portion of the stack, so that the protocol stack carries the information about the path from the first switching device to the n.sup.th switching device.
18. The method according to claim 17, wherein the protocol stack comprises a Multiprotocol Label Switching (MPLS) stack, and one MPLS header of the MPLS stack carries a signpost of one switching device.
19. The method according to claim 11, wherein: the second switching device identifies, according to the signpost of the third switching device, a next-hop switching device of the data is the third switching device, wherein the signpost of the third switching device comprises at least one of the following: a device identifier of the third switching device, an identifier of the port, on the second switching device, of the third switching device, and an identifier of a link between the third switching device and the second switching device.
20. The method according to claim 11, wherein: the first switching device further sends, according to the first forwarding rule, a tenant identifier to the second switching device connected to the first switching device, so as to send the tenant identifier to the n.sup.th switching device by using the second switching device; and the n.sup.th switching device determines the destination end of the data according to the tenant identifier and the characteristic information of the data, and then sends the data to a subsequent switching device.
21. A method for implementing data forwarding, wherein the method is applied to a software-defined networking (SDN) system, the SDN system comprising an SDN control device and n switching devices, a first switching device of the n switching devices being connected to a source end of data, an n.sup.th switching device being connected to a destination end of the data, and n≧3, wherein the method comprises: receiving, by the SDN control device, a request for a forwarding rule from the first switching device, wherein the request carries characteristic information of the data; sending, by the SDN control device, a first forwarding rule to the first switching device according to the characteristic information of the data, wherein the first forwarding rule is configured to instruct the first switching device to send, to a second switching device connected to the first switching device, the data and information about a path from the first switching device to the n.sup.th switching device according to the first forwarding rule; and the SDN control device sends an n.sup.th forwarding rule to the nth switching device according to the characteristic information of the data.
22. The method according to claim 21, wherein: the SDN control device delivers a second forwarding rule to a second of n switching devices according to network topology information of the SDN system, wherein the second forwarding rule is configured to instruct the second switching device to: obtain a port, on the second switching device, of a third switching device by using a signpost of the third switching device in the information about the path from the first switching device to the n.sup.th switching device, and forward the data to the third switching device through the port, on the second switching device.
23. The method according to claim 22, wherein: the SDN control device selects a data forwarding path according to the characteristic information of the data and the network topology information of the SDN system, wherein the information about the path from the first switching device to the n.sup.th switching device comprises signposts of the third switching device to the n.sup.th switching device that are on the data forwarding path; and a match field of the second forwarding rule is the signpost of the third switching device.
24. The method according to claim 23, wherein: the SDN control device delivers an i.sup.th forwarding rule to an i.sup.th switching device according to the network topology information of the SDN system, wherein the i.sup.th forwarding rule is configured to instruct the i.sup.th switching device to obtain a port, on the i.sup.th switching device, of an i+1.sup.th switching device and forward the data to the i+1.sup.th switching device through the port, on the i.sup.th switching device, of the i+1.sup.th switching device, wherein i is an integer greater than or equal to 3, and i is less than n.
25. The method according to claim 24, wherein the information about the path from the i−1.sup.th switching device to the n.sup.th switching device comprises signposts of the i+1.sup.th switching device to the n.sup.th switching device that are on the data forwarding path; and if the i+1.sup.th switching device is not the n.sup.th switching device, the i.sup.th forwarding rule is further configured to instruct the i.sup.th switching device to: remove the signpost of the i+1.sup.th switching device, and send, to the i+1.sup.th switching device, information about a path from the i.sup.th switching device to the n.sup.th switching device, wherein the information about the path from the i.sup.th switching device to the n.sup.th switching device comprises signposts of an i+2.sup.th switching device to the n.sup.th switching device that are on the data forwarding path.
26. The method according to claim 22, wherein: the signpost of the third switching device comprises at least one of the following: a device identifier of the third switching device, an identifier of the port, on the second switching device, of the third switching device, and an identifier of a link between the third switching device and the second switching device.
27. The method according to claim 21, wherein: the first forwarding rule further instructs the first switching device to send a tenant identifier to the second switching device, so as to send the tenant identifier to the n.sup.th switching device via the second switching device; and the i.sup.th forwarding rule instructs the n.sup.th switching device to determine the destination end according to the tenant identifier and the characteristic information of the data.
28. A software-defined networking (SDN) control device, wherein the SDN control device comprises a processor, a memory, and a communications port, wherein: the communications port is configured to communicate with each of a plurality of switching devices in an SDN system, a first switching device of the plurality of switching devices being connected to a source end of data, and an n.sup.th switching device of the plurality of switching devices being connected to a destination end of the data; the memory is configured to store a program; and the processor is configured to execute the program to implement the following process: processing a request for a forwarding rule from the first switching device, wherein the request carries characteristic information of the data and the first switching device; sending a first forwarding rule to the first switching device according to the characteristic information of the data, wherein the first forwarding rule is configured to instruct the first switching device to send, to a second switching device connected to the first switching device, the data and information about a path from the first switching device to the n.sup.th switching device according to the first forwarding rule; and sending an n.sup.th forwarding rule to the n.sup.th switching device according to the characteristic information of the data.
Description
BRIEF DESCRIPTION OF DRAWINGS
[0132] To describe the technical solutions in the embodiments of the present invention more clearly, the following briefly introduces the accompanying drawings required for describing the embodiments or the prior art.
[0133]
[0134]
[0135]
[0136]
[0137]
[0138]
[0139]
[0140]
[0141]
[0142]
[0143]
[0144]
[0145]
[0146]
[0147]
[0148]
[0149]
[0150]
[0151]
[0152]
[0153]
[0154]
[0155]
[0156]
[0157]
[0158]
DESCRIPTION OF EMBODIMENTS
[0159] The following clearly describes the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Apparently, the described embodiments are a part rather than all of the embodiments of the present invention.
[0160] To resolve a problem that performance of devices in SDN is affected because a switching device in a backbone network (which may also be referred to as a switching device at a backbone layer, or may be referred to as a backbone switching device) has too many flow entries that occupy resources of a control device in software-defined networking (Software Defined Networking, SDN) and the switching device at the backbone layer in the SDN, a data forwarding method, device, and system in SDN are provided in the embodiments of the present invention.
[0161] SDN System 1
[0162] According to functions of a switching device, a backbone network of a data center (Data Center, DC) may be divided into three layers: an access layer, an aggregation layer, and a core layer. A virtual layer is applied to a network edge after a DC resource is virtualized.
[0163] A source end virtual host (source end for short) is mounted on a virtual switching device 1, a destination end virtual host (destination end for short) is mounted on a virtual switching device 2, and the two hosts communicate with each other by using the backbone network of the DC. There is a data channel between the source end and the destination end: the virtual switching device 1—an access switching device 1—an aggregation switching device 1—the core switching device—an aggregation switching device 2—an access switching device 2—the virtual switching device 2. A packet is sent from the source end, enters the virtual switching device 1 through a port 2 of the virtual switching device 1, and goes out through a port 5 of the virtual switching device 1; enters the access switching device 1 through a port 7 of the access switching device 1, and goes out through a port 4 of the access switching device 1; enters the aggregation switching device 1 through a port 6 of the aggregation switching device 1, and goes out through a port 3 of the aggregation switching device 1; enters the core switching device through a port 5 of the core switching device, and goes out through a port 9 of the core switching device; enters the aggregation switching device 2 through a port 8 of the aggregation switching device 2, and goes out through a port 7 of the aggregation switching device 2; enters the access switching device 2 through a port 6 of the access switching device 2, and goes out through a port 8 of the access switching device 2; enters the virtual switching device 2 through a port 3 of the virtual switching device 2, and goes out through a port 1 of the virtual switching device 2; and finally reaches the destination end, so as to implement transmission of the data packet.
[0164] The source end or the destination end may be a device such as a mobile terminal (which includes but is not limited to a laptop, a mobile phone, a tablet, a palmtop computer, a POS, an MP3, and a navigator), a fixed terminal (which includes but is not limited to a desktop computer, a printer, a fax machine, a digital projector, and a digital television), or a wireless access terminal. The virtual switching device 1, the access switching device 1, the aggregation switching device 1, the core switching device, the aggregation switching device 2, the access switching device 2, and the virtual switching device 2 are all switching devices in SDN.
[0165] In the prior art, the packet sent from the source end enters the virtual switching device 1; and if no matched flow entry exists, the virtual switching device 1 reports a PacketIn packet (which may also be referred to as a PacketIn event) to a control device. The control device analyzes source information and destination information of the packet, selects a proper forwarding path according to a current network topology, and customizes a flow entry and delivers the flow entry to each switching device on the path. It should be noted that, multiple flow tables can be configured for one switching device, and multiple flow entries can be configured for each flow table. Therefore, with an increase of packets, flow entries of a switching device in the backbone network continuously multiply. A large number of flow entries occupy resources of a control device in the SDN and the switching device at a backbone layer in the SDN, and affect performance of devices in the SDN.
[0166] An SDN control device (which may be also referred to as an SDN control device) in this embodiment of the present invention delivers a part of flow entries (flow entry) according to network topology information, where this part of flow entries may be referred to as proactive flow entries (proactive flow entry). The proactive flow entries are unrelated to a service, but only related to the network topology information. After being delivered, the proactive flow entries may be reused during data forwarding (which may also be referred to as sending). After the packet is received, if no matched flow entry exists, the SDN control device delivers another part of flow entries to some switching devices, where this part of flow entries may be referred to as reactive flow entries (reactive flow entry). The reactive flow entries are related to the service, but are delivered only to partial switching devices. In a case of one-to-one communication, the SDN control device may deliver a reactive flow entry only to a virtual switching device in an edge network, and does not need to deliver a reactive flow entry to a switching device in a backbone network. The virtual switching device encapsulates a forwarding path to a destination end for a packet, and a switching device at a backbone layer executes a forwarding operation according to the forwarding path carried in the packet. Therefore, flow entries delivered by the SDN control device to a switching device in the backbone network are reduced, so that occupation of resources of the SDN control device and the switching device at the backbone layer by flow entries is reduced. Details are described in the following.
[0167] Data Forwarding Method 1
[0168] 1. Proactive Flow Entry
[0169] Before delivering a proactive flow entry according to network topology information, an SDN control device collects the network topology information. For example, the SDN control device monitors statuses of all switching devices, and collects the network topology information according to a connection relationship between the switching devices.
[0170] During a process of collecting the network topology information, the SDN control device needs to allocate signposts to the switching devices. A signpost of a switching device is used by a previous-hop switching device to uniquely identify the switching device, and is also used by the SDN control device to uniquely identify the switching device. It is assumed that the switching device is a first switching device, a previous-hop switching device of the switching device is a second switching device, and a next-hop switching device of the switching device is a third switching device. The second switching device identifies, according to a signpost of the first switching device, that a next-hop switching device of data is the first switching device; similarly, the first switching device identifies, according to a signpost of the third switching device, that a next-hop switching device of the data is the third switching device. No limitation is imposed on a specific implementation form of a signpost in the present invention. The signpost of the first switching device may include but is not limited to: a device identifier of the first switching device, an identifier (that is, a port number) of a port, on the second switching device, of the first switching device, or an identifier of a link between the first switching device and the second switching device. Similarly, the signpost of the third switching device may include but is not limited to: a device identifier of the third switching device, an identifier of a port, on the first switching device, of the third switching device, or an identifier of a link between the third switching device and the first switching device.
[0171] Specifically, as shown in
[0172] For the SDN system shown in
TABLE-US-00001 TABLE 1 Switching devices Device identifiers Virtual switching device 1 1 Access switching device 1 11 Aggregation switching device 1 21 Core switching device 3 Aggregation switching device 2 22 Access switching device 2 12 Virtual switching device 2 2
[0173] The device identifiers allocated by the SDN control device to all the switching devices (which may also be referred to as forwarding devices, or referred to as switches) in a DC network are listed in Table 1. The virtual switching device 1 obtains the device identifier 1, the access switching device 1 obtains the device identifier 11, the aggregation switching device 1 obtains the device identifier 21, the core switching device obtains the device identifier 3, the aggregation switching device 2 obtains the device identifier 22, the access switching device 2 obtains the device identifier 12, and the virtual switching device 2 obtains the device identifier 2.
[0174] Optionally, as shown in
[0175] A procedure in which the SDN control device collects the network topology information by using a signpost allocated to a switching device is shown in
[0176] In the SDN system shown in
TABLE-US-00002 TABLE 2 Switching devices Adjacent switching devices Ports Virtual switching device 1 Access switching device 1 5 Access switching device 1 Virtual switching device 1 7 Aggregation switching device 1 4 Aggregation switching device 1 Access switching device 1 6 Core switching device 3 Core switching device Aggregation switching device 1 5 Aggregation switching device 2 9 Aggregation switching device 2 Core switching device 8 Access switching device 2 7 Access switching device 2 Aggregation switching device 2 6 Virtual switching device 2 8 Virtual switching device 2 Access switching device 2 3
[0177] Specifically, the virtual switching device 1 is connected (which may also be referred to as connected, or may be referred to as adjacent) to the access switching device 1 on the port 5. The access switching device 1 is connected to the virtual switching device 1 and the aggregation switching device 1 respectively on the ports 7 and 4. The aggregation switching device 1 is connected to the access switching device 1 and the core switching device respectively on the ports 6 and 3. The core switching device is connected to the aggregation switching devices 1 and 2 respectively on the ports 5 and 9. The aggregation switching device 2 is connected to the core switching device and the access switching device 2 respectively on the ports 8 and 7. The access switching device 2 is connected to the aggregation switching device 2 and the virtual switching device 2 respectively on the ports 6 and 8. The virtual switching device 2 is connected to the access switching device 2 on the port 3.
[0178] When the SDN control device listens and obtains that a switching device is disconnected because of various reasons, the SDN control device may retrieve a signpost that has been allocated to the switching device as shown in
[0179] In addition, as shown in
[0180] Certainly, when a new switching device accesses the SDN system, the SDN control device also updates the network topology information.
[0181] After the SDN control device allocates the signposts to all the switching devices, and learns the network topology information in
[0182] The proactive flow entries are flow entries that are actively delivered by the SDN control device according to the collected network topology information and are unrelated to a service. As described above, the proactive flow entries are unrelated to a service, but only related to the network topology information. After being delivered, the proactive flow entries may be reused during packet forwarding.
[0183] For the SDN system shown in
TABLE-US-00003 TABLE 3 Switching devices Proactive flow entries Access Match condition: signpost of the virtual switching device switching 1; execution action: removing the signpost of the virtual device 1 switching device 1 and sending the packet to the port 7 Match condition: signpost of the aggregation switching device 1; execution action: removing the signpost of the aggregation switching device 1 and sending the packet to the port 4 Aggregation Match condition: signpost of the access switching device switching 1; execution action: removing the signpost of the access device 1 switching device 1 and sending the packet to the port 6 Match condition: signpost of the core switching device; executionaction: removing the signpost of the core switching device and sending the packet to the port 3 Core switching Match condition: signpost of the aggregation switching device device 1; execution action: removing the signpost of the aggregation switching device 1 and sending the packet to the port 5 Match condition: signpost of the aggregation switching device 2; execution action: removing the signpost of the aggregation switching device 2 and sending the packet to the port 9 Aggregation Match condition: signpost of the core switching device; switching execution action: removing the signpost of the core device 2 switching device and sending the packet to the port 8 Match condition: signpost of the access switching device 2; execution action: removing the signpost of the access switching device 2 and sending the packet to the port 7 Access Match condition: signpost of the aggregation switching switching device 2; execution action: removing the signpost of device 2 the aggregation switching device 2 and sending the packet to the port 6 Match condition: signpost of the virtual switching device 2; execution action: removing the signpost of the virtual switching device 2 and sending the packet to the port 8
[0184] It can be learned from Table 3 that, a proactive flow entry includes two parts, where one part is a match condition (which is also referred to as a match field, and is described as a match field in English), and the other part is an execution action (which is also referred to as an action field, and is described as an action in English). The proactive flow entry delivered by the SDN control device meets the following characteristics:
[0185] a match condition: a signpost of an adjacent switching device; and
[0186] an execution action: removing the signpost of the adjacent switching device, and sending the packet to a port of the adjacent switching device.
[0187] “Adjacent” in this embodiment of the present invention may also be described as “connected” or “connected in an adjacent manner”. For the system architecture diagram shown in
[0188] The proactive flow entries delivered by the SDN control device to all the switching devices in the backbone network are complete proactive flow entries. A packet can be sent from a source end to a destination end by using a part of the flow entries.
[0189] The SDN system in
TABLE-US-00004 TABLE 4 Switching devices Proactive flow entries Access Match condition: signpost of the aggregation switching switching device 1; execution action: removing the signpost of device 1 the aggregation switching device 1 and sending the packet to the port 4 Aggregation Match condition: signpost of the core switching device; switching execution action: removing the signpost of the core device 1 switching device and sending the packet to the port 3 Core switching Match condition: signpost of the aggregation switching device device 2; execution action: removing the signpost of the aggregation switching device 2 and sending the packet to the port 9 Aggregation Match condition: signpost of the access switching switching device 2; execution action: removing the signpost of device 2 the access switching device 2 and sending the packet to the port 7 Access Match condition: signpost of the virtual switching switching device 2; execution action: removing the signpost of device 2 the virtual switching device 2 and sending the packet to the port 8
[0190] On the contrary, if a host connected to the virtual switching device 1 is the destination end, and a host connected to the virtual switching device 2 is the source end, proactive flow entries that need to be used for the packet sent from the source end to the destination end are listed in Table 5. Table 5 is a part of Table 3.
TABLE-US-00005 TABLE 5 Switching devices Proactive flow entries Access Match condition: signpost of the virtual switching device switching 1; execution action: removing the signpost of the virtual device 1 switching device 1 and sending the packet to the port 7 Aggregation Match condition: signpost of the access switching device switching 1; execution action: removing the signpost of the access device 1 switching device 1 and sending the packet to the port 6 Core Match condition: signpost of the aggregation switching switching device 1; execution action: removing the signpost of the device aggregation switching device 1 and sending the packet to the port 5 Aggregation Match condition: signpost of the core switching device; switching execution action: removing the signpost of the core device 2 switching device and sending the packet to the port 8 Access Match condition: signpost of the aggregation switching switching device 2; execution action: removing the signpost of the device 2 aggregation switching device 2 and sending the packet to the port 6
[0191] According to the foregoing description, generally, the SDN control device actively delivers a proactive flow entry to a switching device. The proactive flow entry is used to indicate a match condition and an action of the switching device. Table 3 is used as an example. The proactive flow entry delivered by the SDN control device to the access switching device 1 has two items, where: a match condition in one item of the two items is the signpost of the virtual switching device 1, and an execution action is to remove the signpost of the virtual switching device 1 and send the packet to the port 7; a match condition in the other item is the signpost of the aggregation switching device 1, and an execution action is to remove the signpost of the aggregation switching device 1 and send the packet to the port 4. Similarly, the proactive flow entry delivered by the SDN control device to the aggregation switching device 1 has two items, where: a match condition in one item of the two items is the signpost of the access switching device 1, and an execution action is to remove the signpost of the access switching device 1 and send the packet to the port 6; a match condition in the other item is the signpost of the core switching device, and an execution action is to remove the signpost of the core switching device and send the packet to the port 3. The proactive flow entry delivered by the SDN control device to the core switching device has two items, where: a match condition in one item of the two items is the signpost of the aggregation switching device 1, and an execution action is to remove the signpost of the aggregation switching device 1 and send the packet to the port 5; a match condition in the other item is the signpost of the aggregation switching device 2, and an execution action is to remove the signpost of the aggregation switching device 2 and send the packet to the port 9. The proactive flow entry delivered by the SDN control device to the aggregation switching device 2 has two items, where: a match condition in one item of the two items is the signpost of the core switching device, and an execution action is to remove the signpost of the core switching device and send the packet to the port 8; a match condition in the other item is the signpost of the access switching device 2, and an execution action is to remove the signpost of the access switching device 2 and send the packet to the port 7. The proactive flow entry delivered by the SDN control device to the access switching device 2 also has two items, where: a match condition in one item of the two items is the signpost of the aggregation switching device 2, and an execution action is to remove the signpost of the aggregation switching device 2 and send the packet to the port 6; a match condition in the other item is the signpost of the virtual switching device 2, and an execution action is to remove the signpost of the virtual switching device 2 and send the packet to the port 8.
[0192] In some cases, if the SDN control device receives a PacketIn packet sent by a switching device, the SDN control device checks whether the SDN control device has delivered, to the switching device, a proactive flow entry that arrives at an adjacent switching device. If the SDN control device has not delivered the proactive flow entry to the switching device, the SDN control device delivers, to the switching device, a proactive flow entry that arrives at the adjacent switching device; if the proactive flow entry has been delivered, the SDN control device ignores the PacketIn packet.
[0193] In addition, similar to
[0194] When a switching device is disconnected, the SDN control device finds that the switching device is disconnected; the SDN control device checks whether the disconnected switching device has an adjacent switching device; if the disconnected switching device has an adjacent switching device, the SDN control device deletes a proactive flow entry that is from the adjacent switching device to the disconnected switching device, as listed in Table 3; or if the disconnected switching device does not have an adjacent switching device, the procedure ends.
[0195] When a port status of a switching device changes, the SDN control device finds that the port status of the switching device changes (for example, a port becomes abnormal); the SDN control device checks whether the port whose status changes is connected to an adjacent switching device; and if the port whose status changes is connected to an adjacent switching device, the SDN control device deletes a proactive flow entry that is from the adjacent switching device to the port whose status changes; or if the port whose status changes is not connected to an adjacent switching device, the procedure ends. Specifically, the SDN control device may update Table 2 and Table 3. Details are not described again.
[0196] 2. Reactive Flow Entry
[0197] As shown in
[0198] 101. A source end sends data (which may also be referred to as a packet) to a source edge switching device (for example, the virtual switching device 1) connected to the source end.
[0199] 102-103. The virtual switching device 1 queries a locally stored flow entry, and if a flow entry that matches the data is not found, the virtual switching device 1 requests a forwarding rule of the data from an SDN control device, where the request carries characteristic information of the data.
[0200] It should be noted that, the virtual switching device 1 may send the entire data to the SDN control device, or may send only the characteristic information of the data provided that the request carries the characteristic information of the data. The characteristic information of the data may be any field in a data packet format, and includes but is not limited to at least one of the following: a source IP address, a destination IP address, source MAC, or destination MAC of the data, or a protocol type.
[0201] 104-105. The SDN control device selects a proper forwarding path according to the characteristic information of the data and current network topology information. The SDN control device may select an optimal forwarding path from multiple forwarding paths; for example, the SDN control device may select a shortest forwarding path or a forwarding path that has a lowest traffic weight.
[0202] As shown in
[0203] As shown in
[0204] The SDN control device returns the forwarding rule (that is, a reactive flow entry 1) of the data to the virtual switching device 1 according to the selected forwarding path, as listed in Table 6.
TABLE-US-00006 TABLE 6 Switching device Reactive flow entry Virtual Match condition: characteristic information of the data switching Execution action: encapsulating information about a path device 1 from the virtual switching device 1 to the virtual switching device 2 and sending, to the port 5, the data in which the path information is encapsulated
[0205] Similar to a proactive flow entry, a reactive flow entry also includes two parts, where one part is a match condition (which is also referred to as a match field, and is described as a match field in English), and the other part is an execution action (which is also referred to as an action field, and is described as an action in English). The reactive flow entry is also used to indicate a match condition and an action of a switching device.
[0206] Specifically, the reactive flow entry 1 includes the following information:
[0207] a match condition: the characteristic information of the data; and
[0208] an execution action: encapsulating the information about the path from the virtual switching device 1 to the virtual switching device 2 and sending, to the port 5, the data in which the path information is encapsulated. The virtual switching device 2 is a switching device connected to the destination end of the data.
[0209] Preferably, the information about the path from the virtual switching device 1 to the virtual switching device 2 includes signposts of the aggregation switching device 1 to the virtual switching device 2. An encapsulation process is described in detail in
[0210] 106. In addition, the SDN control device further sends the forwarding rule of the data to the virtual switching device 2, that is, delivers a reactive flow entry 2 to the virtual switching device 2, as listed in Table 7.
TABLE-US-00007 TABLE 7 Switching device Reactive flow entry Virtual Match condition: characteristic information of the data switching Execution action: sending the data to the port 1 device 2
[0211] The reactive flow entry 2 includes the following information:
[0212] a match condition: the characteristic information of the data; and
[0213] an execution action: sending the packet to the port 1, that is, a port connected to the destination end.
[0214] It should be noted that there is no order between 105 and 106.
[0215] After the foregoing process, backbone switching devices receive proactive flow entries from the SDN control device; after an edge switching device receives the data, the edge switching device requests a reactive flow entry from the SDN control device. For details about how to transmit the data from the source end to the destination end according to the proactive flow entries and the reactive flow entry, refer to
[0216] 3. Data Forwarding
[0217]
[0218] 107-108. According to the reactive flow entry 1, the virtual switching device 1 matches the characteristic information of the data, encapsulates the data and the information about the path from the virtual switching device 1 to the virtual switching device 2, and then sends the data and the path information to the access switching device 1 through a port 5 indicated by the reactive flow entry 1.
[0219] Specifically, the virtual switching device 1 may carry the information about the path from the virtual switching device 1 to the virtual switching device 2 by using a protocol stack. The following uses an example in which the protocol stack is a Multiprotocol Label Switching (Multiple Protocol Label Switching, MPLS) stack for description.
[0220] A format of the MPLS protocol is shown in
[0221] One MPLS Header (MPLS header) carries a signpost of one switching device. Multiple MPLS headers are nested sequentially and carry a group of signposts, so as to form a forwarding path. The virtual switching device 1 pushes the information about the path from the virtual switching device 1 to the virtual switching device 2 into the MPLS stack according to the forwarding path selected by the SDN control device. The information about the path from the virtual switching device 1 to the virtual switching device 2 includes the signposts of the aggregation switching device 1 to the virtual switching device 2. The virtual switching device 1 shifts the signpost of the virtual switching device 2 into the bottom of the stack, and shifts the signpost of the aggregation switching device 1 into the top of the stack. The MPLS stack in which the information about the path from the virtual switching device 1 to the virtual switching device 2 is encapsulated is shown in
[0222] As shown in
[0223] 109-110. The access switching device 1 receives the data and the information about the path from the virtual switching device 1 to the virtual switching device 2 (the information about the path from the virtual switching device 1 to the virtual switching device 2 includes signposts of the aggregation switching device 1 to the virtual switching device 2) that are sent by the virtual switching device 1.
[0224] A proactive flow entry (the foregoing proactive flow entry of the access switching device 1 in Table 3) locally stored in the access switching device 1 is queried according to the signpost of the aggregation switching device 1 in the information about the path from the virtual switching device 1 to the virtual switching device 2. If the signpost of the aggregation switching device 1 is matched, actions are executed: removing the signpost of the aggregation switching device 1 and sending, to the port 4, the data and path information from which the signpost is removed. The access switching device 1 forwards, to the aggregation switching device 1 through the port 4, the data and the path information from which the signpost (that is, the signpost of the aggregation switching device 1) is removed.
[0225] The path information from which the signpost (that is, the signpost of the aggregation switching device 1) is removed is shown in
[0226] 111-112. The aggregation switching device 1 receives the data and information about a path from the access switching device 1 to the virtual switching device 2 (the information about the path from the access switching device 1 to the virtual switching device 2 includes signposts of the core switching device to the virtual switching device 2) that are sent by the access switching device 1.
[0227] A proactive flow entry (the foregoing proactive flow entry of the aggregation switching device 1 in Table 3) locally stored in the aggregation switching device 1 is queried according to the signpost of the core switching device in the information about the path from the access switching device 1 to the virtual switching device 2. If the signpost of the core switching device is matched, actions are executed: removing the signpost of the core switching device and sending, to the port 3, the data and path information from which the signpost is removed. The aggregation switching device 1 forwards, to the core switching device through the port 3, the data and the path information from which the signpost (that is, the signpost of the core switching device) is removed.
[0228] The path information from which the signpost (that is, the signpost of the core switching device) is removed is shown in
[0229] 113-114. The core switching device receives the data and information about a path from the aggregation switching device 1 to the virtual switching device 2 (the information about the path from the aggregation switching device 1 to the virtual switching device 2 includes signposts of the aggregation switching device 2 to the virtual switching device 2) that are sent by the aggregation switching device 1.
[0230] A proactive flow entry (the foregoing proactive flow entry of the core switching device in Table 3) locally stored in the core switching device is queried according to the signpost of the aggregation switching device 2 in the information about the path from the aggregation switching device 1 to the virtual switching device 2. If the signpost of the aggregation switching device 2 is matched, actions are executed: removing the signpost of the aggregation switching device 2 and sending, to the port 9, the data and path information from which the signpost is removed. The core switching device forwards, to the aggregation switching device 2 through the port 9, the data and the path information from which the signpost (that is, the signpost of the aggregation switching device 2) is removed.
[0231] The path information from which the signpost (that is, the signpost of the aggregation switching device 2) is removed is shown in
[0232] 115-116. The aggregation switching device 2 receives the data and information about a path from the core switching device to the virtual switching device 2 (the information about the path from the core switching device to the virtual switching device 2 includes signposts of the access switching device 2 to the virtual switching device 2) that are sent by the core switching device.
[0233] A proactive flow entry (the foregoing proactive flow entry of the aggregation switching device 2 in Table 3) locally stored in the aggregation switching device 2 is queried according to the signpost of the access switching device 2 in the information about the path from the core switching device to the virtual switching device 2. If the signpost of the access switching device 2 is matched, actions are executed: removing the signpost of the access switching device 2 and sending, to the port 7, the data and path information from which the signpost is removed. The aggregation switching device 2 forwards, to the access switching device 2 through the port 7, the data and the path information from which the signpost (that is, the signpost of the access switching device 2) is removed.
[0234] The path information from which the signpost (that is, the signpost of the access switching device 2) is removed is shown in
[0235] 117-118. The access switching device 2 receives the data and information about a path from the aggregation switching device 2 to the virtual switching device 2 (the information about the path from the aggregation switching device 2 to the virtual switching device 2 includes the signpost of the virtual switching device 2) that are sent by the aggregation switching device 2.
[0236] A proactive flow entry (the foregoing proactive flow entry of the access switching device 2 in Table 3) locally stored in the access switching device 2 is queried according to the signpost of the virtual switching device 2 in the information about the path from the aggregation switching device 2 to the virtual switching device 2. If the signpost of the virtual switching device 2 is matched, actions are executed: removing the signpost of the virtual switching device 2 and sending the data to the port 8. The access switching device 2 sends the data to the virtual switching device 2 through the port 8.
[0237] 119-120. The SDN control device delivers the reactive flow entry 2 listed in Table 7 to the virtual switching device 2 according to step 106 in
[0238] In this case, the virtual switching devices and the backbone switching devices (for example, the access switching device 1, the aggregation switching device 1, the core switching device, the aggregation switching device 2, and the access switching device 2) transmit the data from the source end to the destination end by using the reactive flow entries and the proactive flow entries respectively.
[0239] In this embodiment of the present invention, an SDN control device delivers a part of flow entries according to network topology information, where this part of flow entries may be referred to as proactive flow entries. The proactive flow entries are unrelated to a service, but only related to the network topology information. After being delivered, the proactive flow entries may be reused during packet forwarding. After a packet is received, if no matched flow entry exists, the SDN control device delivers another part of flow entries to some switching devices, where this part of flow entries may be referred to as reactive flow entries. The reactive flow entries are related to the service, but are delivered only to some switching devices. In a case of one-to-one communication, the SDN control device may deliver a reactive flow entry only to a virtual switching device in an edge network, and does not need to deliver a reactive flow entry to a switching device in a backbone network. The virtual switching device encapsulates a forwarding path to a destination end for a packet, and a switching device at a backbone layer executes a corresponding operation according to a proactive flow entry. Therefore, flow entries delivered by the SDN control device to a switching device in the backbone network are reduced, so that occupation of resources of the SDN control device and the backbone switching device by flow entries is reduced.
[0240] The foregoing description is given by using an example of an SDN system that has seven switching devices. This embodiment of the present invention may be applicable to an SDN system that has n switching devices. The SDN system includes an SDN control device and n switching devices. A first switching device (for example, the virtual switching device 1 in
[0241] 201. The source end sends data to the first switching device, where the first switching device is an edge switching device of the source end.
[0242] 202-203. The first switching device receives the data sent by the source end, and queries a locally stored flow entry; and if a flow entry that matches the data is not found, the first switching device requests a forwarding rule of the data from an SDN control device, where the request carries characteristic information of the data.
[0243] Similar to 102-103 in
[0244] 204. The SDN control device returns a first forwarding rule (for example, the reactive flow entry 1 in the foregoing embodiment) of the data to the first switching device according to the characteristic information of the data, where the first forwarding rule includes information about a path from the first switching device to the n.sup.th switching device.
[0245] Specifically, the SDN control device selects a proper forwarding path according to the characteristic information of the data and current network topology information. The SDN control device returns the first forwarding rule (for example, the reactive flow entry 1 in the foregoing embodiment) of the data to the first switching device according to the selected forwarding path. When n≧3, the first forwarding rule needs to include the information about the path from the first switching device to the n.sup.th switching device; when n<3, the data is directly sent from the source end to the destination end by using two edge switching devices without passing through a backbone switching device, and in this case, there is no need to carry the path information.
[0246] Preferably, the information about the path from the first switching device to the n.sup.th switching device includes signposts of a third switching device to the n.sup.th switching device that are on the data forwarding path. A second switching device (for example, the access switching device 1 in
[0247] Particularly, if n=3, the information about the path from the first switching device to the n.sup.th switching device includes only a signpost of the third switching device. Signposts of the third switching device to the n.sup.th switching device include only the signpost of the third switching device.
[0248] For details, refer to steps 104 and 105 in
[0249] 205. The SDN control device further returns an n.sup.th forwarding rule (for example, the reactive flow entry 2 in the foregoing embodiment) of the data to the n.sup.th switching device according to the characteristic information of the data.
[0250] The n.sup.th forwarding rule is used to instruct the n.sup.th switching device to forward the data to the destination end.
[0251] For details, refer to steps 104 and 106 in
[0252] 206. The first switching device receives the first forwarding rule (for example, the reactive flow entry 1 in the foregoing embodiment) returned by the SDN control device, and sends, to a second switching device connected to the first switching device, the data and the information about the path from the first switching device to the n.sup.th switching device according to the first forwarding rule.
[0253] Specifically, according to the first forwarding rule, the first switching device matches the characteristic information of the data, encapsulates the data and the information about the path from the first switching device to the n.sup.th virtual switching device, and then sends, through the port indicated by the first forwarding rule, the data and the path information to the second switching device connected to the first switching device.
[0254] The information about the path from the first switching device to the n.sup.th virtual switching device includes the signposts of the third switching device to the n.sup.th switching device that are on the data forwarding path. The first switching device successively pushes the signposts of the n.sup.th switching device to the third switching device into a protocol stack, where a signpost of the n.sup.th switching device is shifted into the bottom of the stack, and the signpost of the third switching device is shifted into the top of the stack, so that the protocol stack carries the information about the path from the first switching device to the n.sup.th switching device.
[0255] The protocol stack includes but is not limited to an MPLS stack.
[0256] A signpost of a switching device is similar to that in the foregoing embodiment. For example, the second switching device is configured to identify, according to the signpost of the third switching device, that a next-hop switching device of the data is the third switching device, where the signpost of the third switching device includes but is not limited to at least one of the following: a device identifier of the third switching device, an identifier of a port, on the second switching device, of the third switching device, or an identifier of a link between the third switching device and the second switching device.
[0257] For details, refer to step 107-108 in
[0258] 207. The second switching device receives the data and the information about the path from the first switching device to the n.sup.th switching device that are sent by the first switching device; queries, according to a signpost of a third switching device in the information about the path from the first switching device to the n.sup.th switching device, a second forwarding rule (the proactive flow entry of the access switching device 1 in Table 3) stored in the second switching device; and according to the second forwarding rule, obtains a port, on the second switching device, of the third switching device, and forwards the data to the third switching device through the port, on the second switching device, of the third switching device, where the third switching device is a switching device connected to the second switching device.
[0259] The second forwarding rule is a proactive flow entry, and is unrelated to the data, but is only related to the network topology information. Therefore, a match field of the second forwarding rule includes the signpost of the third switching device rather than the characteristic information of the data. An action field of the second forwarding rule is: removing the signpost of the third switching device and sending, to the port, on the second switching device, of the third switching device, the data from which the signpost is removed.
[0260] If n≧4, the second switching device further sends, to the third switching device, information about a path from the second switching device to the n.sup.th switching device, where the information about the path from the second switching device to the n.sup.th switching device includes signposts of the fourth switching device to the n.sup.th switching device that are on the data forwarding path, and the fourth switching device (for example, the core switching device in
[0261] Particularly, if n=4, the information about the path from the second switching device to the n.sup.th switching device includes only a signpost of the fourth switching device. Signposts of the fourth switching device to the n.sup.th switching device include only the signpost of the fourth switching device.
[0262] For details, refer to step 109-110 in
[0263] According to the foregoing method, information about a path from a first switching device to an n.sup.th switching device is determined, the first switching device encapsulates the information about the path from the first switching device to the nth switching device and sends the information to a second switching device, and the second switching device forwards the data according to the information about the path from the first switching device to the nth switching device and a second forwarding rule, where the second forwarding rule is only related to network topology information, but unrelated to the data (or packet). Therefore, the second switching device does not need to acquire a forwarding rule each time after receiving data, that is, there is no need to acquire, each time after receiving data, a flow entry from an SDN control device to perform data forwarding, which reduces consumption of resources used by the SDN control device to deliver a flow entry, and improves utilization of a device.
[0264] In an optional implementation manner, the second forwarding rule is generated when the second switching device establishes a topology relationship with a switching device connected to the second switching device or when a topology relationship changes. When the topology relationship between the second switching device and the switching device connected to the second switching device changes, a forwarding rule between the second switching device and the switching device connected to the second switching device changes accordingly. In this case, when the second switching device establishes a topology relationship with a switching device connected to the second switching device or when a topology relationship changes, the SDN control device generates the second forwarding rule and sends the second forwarding rule to the second switching device, so that the second switching device can perform forwarding according to a latest network topology when performing data forwarding according to the second forwarding rule, which reduces resource consumption of a device and improves accuracy of data forwarding.
[0265] 208. An i.sup.th switching device receives the data and information about a path from an i−1.sup.th switching device to the n.sup.th switching device that are sent by the i−1.sup.th switching device, where the information about the path from the i−1.sup.th switching device to the n.sup.th switching device includes signposts of an i+1.sup.th switching device to the n.sup.th switching device that are on the data forwarding path, i is an integer greater than or equal to 3, and i is less than n.
[0266] The i.sup.th switching device queries, according to a signpost of the i+1.sup.th switching device, an i.sup.th forwarding rule (for example, a proactive flow entry in Table 3) stored in the i.sup.th switching device; and according to the i.sup.th forwarding rule, determines a port, on the it.sup.h switching device, of the i+1.sup.th switching device, and forwards the data to the i+1.sup.th switching device through the port.
[0267] If the i+1.sup.th switching device is not the last switching device on the data forwarding path, that is, not the n.sup.th switching device, the i.sup.th switching device further sends, to the i+1.sup.th switching device, information about a path from the i.sup.th switching device to the n.sup.th switching device according to the i.sup.th forwarding rule (for example, the proactive flow entry in Table 3), where the information about the path from the i.sup.th switching device to the n.sup.th switching device includes signposts of an i+2.sup.th switching device to the n.sup.th switching device that are on the data forwarding path, and the i+2.sup.th switching device is a switching device connected to the i+1.sup.th switching device.
[0268] Specifically, that the i.sup.th switching device sends, to the i+1.sup.th switching device, information about a path from the i.sup.th switching device to the n.sup.th switching device according to the i.sup.th forwarding rule (for example, the proactive flow entry in Table 3) specifically includes: the i.sup.th switching device removes the signpost of the i+1.sup.th switching device from the information about the path from the i−1.sup.th switching device to the n.sup.th switching device according to the i.sup.th forwarding rule to obtain the signposts of the i+2.sup.th switching device to the n.sup.th switching device, and sends, to the i+1.sup.th switching device, the signposts of the i+2.sup.th switching device to the n.sup.th switching device. In addition, the i.sup.th switching device further sends, to the i+1.sup.th switching device, the information about the path from the i.sup.th switching device to the n.sup.th switching device according to the port, on the i.sup.th switching device, of the i+1.sup.th switching device.
[0269] The i.sup.th switching device is, for example, the aggregation switching device 1, the core switching device, the aggregation switching device 2, or the access switching device 2 in
[0270] Step 208 is repeatedly executed until the data is transmitted to the n.sup.th switching device.
[0271] For details, refer to steps 111 to 118 in
[0272] 209. The n.sup.th switching device forwards, if the data received from an n−1.sup.th switching device matches a match field of the n.sup.th forwarding rule, the data to the destination end according to the n.sup.th forwarding rule received from the SDN control device.
[0273] For details, refer to step 119-120 in
[0274] According to the foregoing method for implementing data forwarding, after receiving a forwarding rule (for example, a proactive flow entry in the foregoing embodiment), a backbone switching device does not need to acquire a forwarding rule each time data is being forwarded, which reduces resource overheads of a control device and a switching device, and improves a utilization rate of resources in an SDN system. For example, after receiving a second forwarding rule (which may also be referred to as a flow entry), the backbone switching device (for example, a second switching device) does not need to acquire a forwarding rule each time data is being forwarded. Accordingly, an SDN control device does not need to deliver a second forwarding rule to the second switching device each time the data is being forwarded, that is, the second forwarding rule may be reused during subsequent data forwarding. This reduces a quantity of forwarding rules delivered by the SDN control device to the second switching device, and accordingly reduces a device resource occupation rate of the SDN control device; and network bandwidth between the SDN control device and the second switching device is also reduced accordingly. In addition, the second switching device does not need to receive a second forwarding rule each time the data is being forwarded, which improves a resource utilization rate of the second switching device.
[0275] After a backbone switching device in a DC network receives a proactive flow entry delivered by the SDN control device and an edge switching device receives a reactive flow entry delivered by the SDN control device, packet forwarding can be implemented. Because the backbone switching device receives the proactive flow entry, and there is no need to deliver a flow entry to the backbone switching device each time the data is being forwarded, that is, the proactive flow entry may be reused in subsequent data forwarding, a quantity of flow entries delivered to the backbone switching device is reduced, a device resource occupation rate of the SDN control device is reduced accordingly, and network bandwidth between the SDN control device and the backbone switching device is also reduced accordingly. In addition, the backbone switching device does not need to receive a flow entry each time the data is being forwarded, which improves a resource utilization rate of the backbone switching device.
[0276] In the foregoing data forwarding method 1, there is one source end and one destination end, and the method may be referred to as a unicast solution. The foregoing data forwarding method may be applied to various scenarios, for example, network security, a service, multicast, or broadcast (a scenario of multiple destination ends). Descriptions are given separately as follows:
[0277] SDN System 2
[0278] A tenant may lease a computing resource, a storage resource, or a network resource from a data center (DC) supplier. A user may use the resource leased by the tenant. Therefore, the tenant may deploy various services on a host, in the DC, to which the tenant belongs, and open the services to the user of the tenant.
[0279] The DC allows the tenant to customize a policy to instruct data to shuttle between different services.
[0280] As shown in
[0281] A host that provides various services (or a host on which a service is deployed by the tenant) may be referred to as a service device. As shown in
[0282] The tenant provides a related service by using a host on which the related service is deployed. For example, a host on which a firewall is deployed provides a firewall service, where the host that provides the firewall service is a service device. The following describes a data forwarding method in which the services are added to the SDN system.
[0283] Data Forwarding Method 2
[0284] 1. Proactive Flow Entry
[0285] Similar to the proactive flow entries in the data forwarding method 1, an SDN control device needs to deliver proactive flow entries to all backbone switching devices according to network topology information. For details, refer to the description in the data forwarding method 1, and details are not described again.
[0286] In
[0287] A tenant may request, by using a UI (User Interface, user interface), the SDN control device to add or delete a service device. The tenant initiates, to the SDN control device by using the UI, a request of creating a service device; the SDN control device creates service devices, and the SDN control device allocates identifiers of the service devices to the service devices (refer to Table 8). An identifier of a service device is unique, and is used to distinguish from another service device in a DC network or another type of object that can have a device identifier; the identifier may be independent, or may be combined with another device identifier. An identifier of a service device is used to determine a data forwarding path. The SDN control device makes, by using the UI, a response to the request of creating a service device.
[0288] For the SDN system in
TABLE-US-00008 TABLE 8 Identifiers of the Service devices service devices Host on which a firewall locates 5 Host on which an authentication 6 service locates Host on which a load balancing 7 service locates Host on which a database 1 locates 8 Host on which a database 2 locates 9
[0289] The host on which a firewall locates (which may also be referred to as a firewall service device) obtains an identifier 5; the host on which an authentication service locates (which may also be referred to as an authentication service device) obtains an identifier 6; the host on which a load balancing service locates (which may also be referred to as a load balancing service device) obtains an identifier 7; the host on which a database 1 locates (which may also be referred to as a database 1 service device) obtains an identifier 8; the host on which a database 2 locates (which may also be referred to as a database 2 service device) obtains an identifier 9.
[0290] When the tenant needs to delete a service device, the tenant initiates, to the SDN control device by using the UI, a request of deleting a service device; the SDN control device deletes the service device, and then retrieves an identifier that has been allocated to the service device. The retrieved identifier of the service device may be reallocated to another service device or another type of object that can have a device identifier. The SDN control device makes, by using the UI, a response to the request of deleting a service device.
[0291] The tenant may set a service for a user of the tenant, for example, only the firewall is passed through, or only authentication is performed. Certainly, alternatively, the tenant may set multiple services for the user of the tenant, where there is order between these services. A service sequence that includes multiple services may also be referred to as a service chain. Order, specified by the tenant, of the multiple services indicates a policy on providing a service to the outside by the tenant. A DC allows the tenant to customize different service chains, and then leads data to different service devices according to characteristic information of the data of the user, so as to traverse all service devices on the service chains.
[0292] An operation of adding a service chain by the tenant is also implemented by using the UI. A specific process is as follows: The tenant initiates, to the SDN control device by using the UI, a request of creating a service chain; the SDN control device creates service chains (as listed in Table 9); and the SDN control device returns, to the tenant by using the UI, a result of creating the service chains.
[0293] The service chains created by the SDN control device are listed in Table 9. Certainly, Table 9 is merely exemplary.
TABLE-US-00009 TABLE 9 Service Identifier sequences chains Order of services of service nodes 1 Firewall—Authentication—Database 1 5-6-8 2 Firewall—Authentication—Database 2 5-6-9 3 Firewall—Authentication—Load balancing 5-6-7
[0294] The tenant specifies the database 1 as a data source for the user by using the first service chain; data passes through the firewall, then is authenticated, and finally arrives at the database 1; an identifier sequence of corresponding service devices is 5-6-8. The tenant specifies the database 2 as a data source for the user by using the second service chain; data passes through the firewall, then is authenticated, and finally arrives at the database 2; an identifier sequence of corresponding service devices is 5-6-9. A load balancing service specifies a data source for the user by using the third service chain; data passes through the firewall, then is authenticated, scheduled by the load balancing service, and finally arrives at a database specified by the load balancing service; an identifier sequence of corresponding service devices is 5-6-7.
[0295] After allocating the identifiers of the service devices and the service chains, the SDN control device needs to deliver proactive flow entries to switching devices connected to the service devices. The proactive flow entries delivered to the switching devices connected to the service devices meet the following characteristics:
[0296] a match condition: an identifier of a service device; and
[0297] an execution action: removing the identifier of the service device and sending the data to a service device port.
TABLE-US-00010 TABLE 10 Switching devices Proactive flow entries Access Match condition: identifier of the firewall service device; switching execution action: removing the identifier of the firewall device 1 service device and sending an original packet to a port 1 Access Match condition: identifier of the authentication service switching device; execution action: removing the identifier of the device 3 authentication service device and sending an original packet to a port 8 Access Match condition: identifier of the load balancing service switching device; execution action: removing the identifier of the device 2 load balancing service device and sending an original packet to a port 9 Virtual Match condition: identifier of the database 1 service switching device; execution action: removing the identifier of the device 2 database 1 service device and sending an original packet to a port 6 Match condition: identifier of the database 2 service device; execution action: removing the identifier of the database 2 service device and sending an original packet to a port 1
[0298] The proactive flow entries delivered by the SDN control device to the switching devices connected to the service devices are listed in Table 10. The SDN control device instructs the access switching device 1 to: match a packet that carries the identifier of the firewall service device, remove the identifier of the firewall service device from the packet, and send the original packet to the port 1; instructs the access switching device 3 to: match a packet that carries the identifier of the authentication service device, remove the identifier of the authentication service device from the packet, and send the original packet to the port 8; instructs the access switching device 2 to: match a packet that carries the identifier of the load balancing service device, remove the identifier of the load balancing service device from the packet, and send the original packet to the port 9; instructs the virtual switching device 2 to: match a packet that carries an identifier of a database service device, remove the identifier of the database service device from the packet, and send the original packet to a corresponding port (if the database service device is the database 1 service device, the original packet is sent to the port 6; if the database service device is the database 2 service device, the original packet is sent to the port 1).
[0299] Certainly, the tenant may also delete a service chain. A specific process is as follows: The tenant initiates, to the SDN control device by using the UI, a request of deleting a service chain; the SDN control device deletes the service chain; and the SDN control device returns, to the tenant by using the UI, a result of deleting the service chain. In addition, the SDN control device further needs to delete a proactive flow entry of a switching device connected to a service device.
[0300] 2. Reactive Flow Entry
[0301] 301. A user sends data (which may also be referred to as a packet, or a data flow, or a user flow) to a switching device (for example, a virtual switching device 1) connected to the user.
[0302] 302-303. The switching device (for example, the virtual switching device 1) connected to the user queries a locally stored flow entry, and if a flow entry that matches the data is not found, the virtual switching device 1 requests a forwarding rule of the data from an SDN control device, where the request carries characteristic information of the data.
[0303] It should be noted that, the virtual switching device 1 may send the entire data to the SDN control device, or may send only the characteristic information of the data provided that the request carries the characteristic information of the data. The characteristic information of the data may be any field in a data packet format, and includes but is not limited to at least one of the following: a source IP address, a destination IP address, source MAC, or destination MAC of the data, or a protocol type.
[0304] 304-305. The SDN control device selects a proper forwarding path according to the characteristic information of the data and current network topology information and a service policy of a tenant. The SDN control device may select an optimal forwarding path from multiple forwarding paths; for example, the SDN control device may select a shortest forwarding path or a forwarding path that has a lowest traffic weight.
[0305] Reference may be specifically made to the description in
[0306] The SDN control device delivers, according to the selected forwarding path, the forwarding rule (that is, a reactive flow entry 1′) of the data to the switching device connected to the user, as listed in Table 11.
TABLE-US-00011 TABLE 11 Switching device Reactive flow entry Virtual Match condition: characteristic information of the data; switching execution action: encapsulating the identifier of the device 1 firewall service device and sending the packet in which the identifier is encapsulated to a port 5
[0307] Similar to a proactive flow entry, a reactive flow entry (flow entry) also includes two parts, where one part is a match condition (which is also referred to as a match field, and is described as a match field in English), and the other part is an execution action (which is also referred to as an action field, and is described as an action in English).
[0308] Specifically, the reactive flow entry includes the following information:
[0309] a match condition: the characteristic information of the data; and
[0310] an execution action: encapsulating the identifier of the firewall service device, and sending the packet in which the identifier is encapsulated to the port 5.
[0311] Because the virtual switching device 1 is connected to the access switching device 1, and the access switching device 1 is connected to the firewall service device, only the identifier of the firewall service device is encapsulated into the packet sent by the virtual switching device 1 to the access switching device 1, and there is no need to encapsulate information about a path to the access switching device 1.
[0312] 306. In addition, the SDN control device further needs to deliver forwarding rules (that is, reactive flow entries 2′) of the data to switching devices connected to service devices, as listed in Table 12.
TABLE-US-00012 TABLE 12 Switching devices Reactive flow entries Access Match condition: characteristic information of the switching processed data; execution action: encapsulating the device 1 identifier of the authentication service device and information about a path from the access switching device 1 to a switching device connected to the authentication service device, and sending the data in which the identifier and the path information are encapsulated to a port 4 Access Match condition: characteristic information of the switching processed data; execution action: encapsulating the device 3 identifier of the load balancing service device and information about a path from the access switching device 3 to a switching device connected to the load balancing service device, and sending the data in which the identifier and the path information are encapsulated to a port 5 Access Match condition: characteristic information of the switching processed data; execution action: encapsulating the device 2 identifier of the database service device, and sending the data in which the identifier is encapsulated to a port 8
[0313] An identifier of a next-hop service device is encapsulated for the data according to locations of the service devices on a service chain; in some cases, information about a path to a next-hop service device needs to be further encapsulated, and then the data in which the identifier and the path information are encapsulated is sent to an egress port of a switching device.
[0314] The reactive flow entries 2′ delivered by the SDN control device to the switching devices connected to the service devices are listed in Table 12. The reactive flow entries 2′ instruct the access switching device 1 to: match the characteristic information of the processed data, encapsulate the identifier of the authentication service device and the information about the path from the access switching device 1 to the switching device connected to the authentication service device, and then send the data in which the identifier and the path information are encapsulated to the port 4; instruct the access switching device 3 to: match the characteristic information of the processed data, encapsulate the identifier of the load balancing service device and the information about the path from the access switching device 3 to the switching device connected to the load balancing service device, and then send the data in which the identifier and the path information are encapsulated to the port 5; and instruct the access switching device 2 to: match the characteristic information of the processed data, encapsulate the identifier of the database service device, and then send the data in which the identifier is encapsulated to the port 8.
[0315] It should be noted that, because the database is a last-hop service on the service chain, the SDN control device does not need to deliver a reactive flow entry to the virtual switching device 2 connected to the database, and only need to deliver the proactive flow entries listed in Table 10. There is no order between 305 and 306.
[0316] After the foregoing process, the backbone switching devices receive the proactive flow entries from the SDN control device, and the switching devices connected to the service devices receive other proactive flow entries; after the data is received, the switching device connected to the user requests a reactive flow entry from the SDN control device, and the switching devices connected to the service devices also receive other reactive flow entries. For details about how to transmit the data from the user to the database according to the proactive flow entries and the reactive flow entries, refer to
[0317] In the foregoing embodiment, a proactive flow entry and a reactive flow entry are delivered in a scenario in which different service devices exist, which implements that a flow entry does not need to be repeatedly delivered when a tenant-customized policy is met to implement flexible shuttling of data or a service flow between different services. Therefore, device resource occupation rates of an SDN control device and a switching device are reduced, and performance and efficiency of the SDN control device and the switching device are also improved accordingly.
[0318] 3. Data Forwarding
[0319] It should be noted that, because data that enters a service device needs to be original data, when delivering a forwarding path, an SDN control device cannot customize a forwarding path of an entire service chain at a time, and needs to divide the path into multiple segments according to each service device. A start of each segment is a previous service device, an end is a next service device, and a path delivery manner of each segment is similar to an implementation manner shown in the data forwarding method 1. Details are described in the following with reference to
[0320]
[0321] 407-408. According to the reactive flow entry 1′ in Table 11, a switching device connected to the user (for example, the virtual switching device 1) matches the characteristic information of the data, encapsulates an identifier of a firewall service device, and then sends the data and the identifier of the firewall service device to the access switching device 1 through a port 5 indicated by the reactive flow entry 1′.
[0322] 409-410. The access switching device 1 receives the identifier of the firewall service device sent by the virtual switching device 1. According to the proactive flow entry of the access switching device 1 in Table 10, the access switching device 1 matches the identifier of the firewall service device, and then executes actions of removing the identifier of the firewall service device and sending the data to a port 1, where the data arrives at the firewall service device through the port 1, and removing may also be understood as deleting.
[0323] 411-413. The access switching device 1 receives, though the port 1, data processed by a firewall, where the data processed by the firewall is also referred to as data processed by the firewall service device. If the data processed by the firewall matches characteristic information of the processed data, the access switching device 1 executes, according to the reactive flow entry 2′ of the access switching device 1 indicated in Table 12, actions of encapsulating an identifier of a next service device (for example, an identifier of an authentication service device) and information about a path from the access switching device 1 to a switching device (for example, the access switching device 3) connected to the authentication service device and sending, to an aggregation switching device 1 through a port 4, data in which the identifier and the path information are encapsulated.
[0324] The information about the path from the access switching device 1 to the access switching device 3 is similar to the path information in the data forwarding method 1. It is only needed to treat the user as the source end, treat the access switching device 1 as the switching device connected to the source end in the data forwarding method 1, treat the authentication service device as the destination end, and treat the access switching device 3 as the switching device connected to the destination end in the data forwarding method 1. In the SDN system 2, an encapsulation action of the access switching device 1 is similar to that of the virtual switching device 1 in the data forwarding method 1.
[0325] Specifically, the information about the path from the access switching device 1 to the switching device connected to the authentication service device includes signposts of a core switching device to the access switching device 3. The access switching device 1 may carry, by using a protocol stack, the information about the path from the access switching device 1 to the switching device connected to the authentication service device. The following uses an example in which the protocol stack is a Multiprotocol Label Switching (Multiple Protocol Label Switching, MPLS) stack for description. Similar to the data forwarding method 1, one MPLS Header (MPLS header) carries a signpost of one switching device. Multiple MPLS headers are nested sequentially and carry a group of signposts, so as to form a forwarding path. The access switching device 1 pushes the information about the path from the access switching device 1 to the switching device connected to the authentication service device into the MPLS stack according to the forwarding path selected by the SDN control device.
[0326] The information about the path from the access switching device 1 to the switching device connected to the authentication service device includes the signposts of the core switching device to the access switching device 3. The access switching device 1 shifts the identifier of the authentication service device into the bottom of the stack, shifts a signpost of the access switching device 3 into an upper layer of the bottom of the stack, and shifts a signpost of the core switching device into the top of the stack. An MPLS stack in which the identifier of the authentication service device and the information about the path from the access switching device 1 to the switching device connected to the authentication service device are encapsulated is shown in
[0327] As shown in
[0328] 414-415. All backbone switching devices (for example, the aggregation switching device 1, the core switching device, and the aggregation switching device 2) receive the data processed by the firewall service device and the path information, and execute, according to the proactive flow entries listed in Table 3, actions of removing a signpost at an uppermost layer and sending the data to a next-hop switching device. For details, refer to the description in the data forwarding method 1, and details are not described again.
[0329] 416-417. The access switching device 3 receives the data processed by the firewall and the identifier of the authentication service device. According to the proactive flow entry of the access switching device 3 in Table 10, the access switching device 3 matches the identifier of the authentication service device, and then executes actions of removing the identifier of the authentication service device and sending the data to a port 8. The data arrives at the authentication service device through the port 8.
[0330] 418-420. The access switching device 3 receives, though the port 8, data on which authentication processing is performed, where the data on which authentication processing is performed is also referred to as data processed by the authentication service device. If the data on which authentication processing is performed matches characteristic information of the processed data, the access switching device 3 executes, according to the reactive flow entry 2′ of the access switching device 3 listed in Table 12, actions of encapsulating an identifier of a next-hop service device (for example, an identifier of a load balancing service device) and information about a path from the access switching device 3 to a switching device (for example, the access switching device 2) connected to the load balancing service device and sending, to the aggregation switching device 2 through the port 5, data in which the identifier and the path information are encapsulated.
[0331] The information about the path from the access switching device 3 to the switching device (for example, the access switching device 2) connected to the load balancing service device is similar to that in the data forwarding method 1. It is only needed to treat the access switching device 3 as the switching device connected to the source end in the data forwarding method 1, treat the load balancing service device as the destination end, and treat the access switching device 2 as the switching device connected to the destination end. In the SDN system 2, an encapsulation action of the access switching device 3 is similar to that of the virtual switching device 1 in the data forwarding method 1.
[0332] Specifically, the information about the path from the access switching device 3 to the switching device (for example, the access switching device 2) connected to the load balancing service device includes a signpost of the access switching device 2. The access switching device 3 may carry, by using a protocol stack, the information about the path from the access switching device 3 to the switching device (for example, the access switching device 2) connected to the load balancing service device. The access switching device 3 pushes the information about the path from the access switching device 3 to the switching device (for example, the access switching device 2) connected to the load balancing service device into the MPLS stack according to the forwarding path selected by the SDN control device.
[0333] The information about the path from the access switching device 3 to the switching device (for example, the access switching device 2) connected to the load balancing service device includes the signpost of the access switching device 2. The access switching device 3 shifts the identifier of the load balancing service device into the bottom of the stack, shifts the signpost of the access switching device 2 into an upper layer of the bottom of the stack. An MPLS stack in which the identifier of the load balancing service device and the information about the path from the access switching device 3 to the access switching device 2 are encapsulated is shown in
[0334] As shown in
[0335] 421-422. A backbone switching device (for example, the aggregation switching device 2) executes, according to a proactive flow entry listed in Table 3, actions of removing a signpost at an uppermost layer and sending the data to a next-hop switching device. For details, refer to the description in the data forwarding method 1, and details are not described again.
[0336] 423-424. The access switching device 2 receives the data and the identifier of the load balancing service device. According to the proactive flow entry of the access switching device 2 in Table 10, the access switching device 2 matches the identifier of the load balancing service device, and then executes actions of removing the identifier of the load balancing service device and sending the data to a port 9. The data processed by the authentication service device arrives at the load balancing service device through the port 9. It is assumed that the load balancing service device selects, according to a preset policy, a database 1 to perform data processing.
[0337] 425-427. The access switching device 2 receives, though the port 9, data on which load balancing processing is performed, where the data on which load balancing processing is performed is also referred to as data processed by the load balancing service device. If the data processed by the load balancing service device matches characteristic information of the processed data, the access switching device 2 executes, according to the reactive flow entry 2′ of the access switching device 2 listed in Table 12, actions of encapsulating an identifier of a next-hop service device (for example, an identifier of a database 1 service device) and sending, to the virtual switching device 2 through the port 8, data in which the identifier is encapsulated.
[0338] Because an encapsulation operation for a forwarding path is executed on the access switching device 2, the access switching device 2 learns, according to the reactive flow entry 2′ listed in Table 12, that a next-hop switching device of the packet is the virtual switching device 2, and the virtual switching device 2 is a switching device connected to the database service device; path information does not need to be encapsulated.
[0339] 428-429. The virtual switching device 2 receives the data on which load balancing is performed and the identifier of the database 1 service device. According to the proactive flow entry of the virtual switching device 2 in Table 10, the virtual switching device 2 matches the identifier of the database 1 service device, and then executes actions of removing the identifier of the database 1 service device and sending the data to a port 6. The data arrives at the database 1 through the port 6.
[0340] 430. The database 1 processes the data, for example, performing querying, deleting, and adding on the data. Because the database 1 is the last-hop service on the entire service chain, the database 1 may also be referred to as a destination service, and a host on which the database 1 locates may also be referred to as a destination service device.
[0341] When being returned to the user, the data processed by the database 1 does not need to pass through all the service devices, and may be returned to the user only by using an edge switching device and a backbone switching device, which is similar to the data forwarding method 1.
[0342] In this embodiment of the present invention, an SDN control device delivers a part of flow entries according to network topology information, where this part of flow entries may be referred to as proactive flow entries. The proactive flow entries are unrelated to a service, but only related to the network topology information. After being delivered, the proactive flow entries may be reused during packet forwarding. After a packet is received, if no matched flow entry exists, the SDN control device delivers another part of flow entries to some switching devices, where this part of flow entries may be referred to as reactive flow entries. The reactive flow entries are related to the service, but are delivered only to some switching devices. In a case of a service chain, the SDN control device may deliver reactive flow entries only to a switching device connected to a user and a switching device connected to a service device, and does not need to deliver a reactive flow entry to another switching device in a backbone network. The switching device connected to the user and the switching device connected to the service device encapsulate forwarding paths to switching devices connected to next-hop service devices, and a switching device at a backbone layer executes a corresponding operation according to a proactive flow entry. Therefore, flow entries delivered by the SDN control device to a switching device in the backbone network are reduced, so that occupation of resources of the SDN control device and the backbone switching device by flow entries is reduced.
[0343] The foregoing description is given by using an example of an SDN system that has seven switching devices and four service devices. This embodiment of the present invention may be applicable to an SDN system that has any quantity of switching devices and any quantity of service devices.
[0344] If there is only one service device, the switching device connected to the user encapsulates an identifier of the service device, and other steps are similar to that in the data forwarding method 1. In addition to path information, the identifier of the service device further needs to be transmitted. After receiving the identifier of the service device and data, a switching device connected to the service device deletes the identifier of the service device and sends the data to the service device for processing.
[0345] That there are at least two service devices is used as an example for description in the following. An SDN system includes an SDN control device and at least two service devices. The at least two service devices include a first service device and a second service device. The first service device is connected to a first switching device, and the second service device is connected to a second switching device. For a specific data forwarding procedure, refer to
[0346] 500. The SDN control device delivers, according to network topology information of the SDN system, a first forwarding rule (for example, a proactive flow entry listed in Table 10) to the first switching device connected to the first service device.
[0347] Delivery may also be described as returning, or may be described as sending.
[0348] 501. The SDN control device delivers, according to the network topology information of the SDN system, a third forwarding rule (for example, a proactive flow entry listed in Table 10) to the second switching device connected to the second service device.
[0349] 502. If the first switching device is not connected to the second switching device, that is, there is another backbone switching device between the first switching device and the second switching device, according to the network topology information of the SDN system, the SDN control device further delivers a fourth forwarding rule (for example, a proactive flow entry listed in Table 5) to a third switching device, and delivers an i+1.sup.th forwarding rule (for example, a proactive flow entry listed in Table 5) to an i.sup.th switching device, where i is a natural number greater than or equal to 4.
[0350] 503. After receiving data sent by a user, if a switching device connected to the user does not find a matched flow entry locally, the switching device connected to the user requests a forwarding rule of the data from the SDN control device. The SDN control device delivers a reactive flow entry (as listed in Table 11) to the switching device connected to the user, and delivers a reactive flow entry (as listed in Table 12) to a switching device connected to a service device. It is assumed that a reactive flow entry delivered to the first switching device is referred to as a second forwarding rule.
[0351] 601-602. The first switching device receives the data and an identifier of the first service device; and according to the first forwarding rule received from the SDN control device, deletes the identifier of the first service device and sends the data to the first service device for processing.
[0352] If the data does not pass through a service device before passing through the first switching device, the data is original data sent by the switching device connected to the user; if the data passes through at least one service device before passing through the first switching device, the data is data processed by a previous-hop service device.
[0353] 603. Receive the data processed by the first service device.
[0354] When the first switching device is connected to the second switching device, step 604a is executed.
[0355] When the first switching device is not connected to the second switching device, steps 604b to 608b are executed.
[0356] 604a. The first switching device sends, according to the second forwarding rule received from the SDN control device, the data processed by the first service device and an identifier of the second service device to the second switching device.
[0357] 604b. The first switching device sends, to the third switching device according to the second forwarding rule received from the SDN control device, the data processed by the first service device, an identifier of the second service device, and information about a path from the first switching device to the second switching device, where the third switching device is a switching device connected to the first switching device.
[0358] Specifically, the information about the path from the first switching device to the second switching device includes signposts of a fourth switching device to the second switching device that are on a forwarding path of the data processed by the first service device, where the fourth switching device is a switching device connected to the third switching device.
[0359] The first switching device successively pushes, into a protocol stack, the identifier of the second service device and the signposts of the second switching device to the fourth switching device that are on the forwarding path of the data processed by the first service device, where the identifier of the second service device is shifted to the bottom of the stack, a signpost of the second switching device is shifted into an upper layer of the bottom of the stack, and a signpost of the fourth switching device is shifted to the top of the stack. Encapsulation of the path information by the first switching device is similar to that by the virtual switching device 1 in the data forwarding method 1. For details, refer to the description in the data forwarding method 1, and it is only needed to encapsulate the identifier of the second service device into the bottom of the stack.
[0360] 605b-606b. The third switching device is configured to: query, according to a signpost of the fourth switching device, a fourth forwarding rule stored in the third switching device; and according to the fourth forwarding rule, determine a port, on the third switching device, of the fourth switching device, and forward the data processed by the first service device to the fourth switching device through the port, on the third switching device, of the fourth switching device.
[0361] If the fourth switching device is not the second switching device, the third switching device is further configured to: remove the signpost of the fourth switching device from the information about the path from the first switching device to the second switching device to obtain information about a path from the third switching device to the second switching device, and send, to the fourth switching device, the information about the path from the third switching device to the second switching device, according to the fourth forwarding rule.
[0362] The third switching device is configured to identify, according to the signpost of the fourth switching device, that a next-hop switching device of the data is the fourth switching device, where the signpost of the fourth switching device includes at least one of the following: a device identifier of the fourth switching device, an identifier of an egress port, on the third switching device, of the fourth switching device, or an identifier of a link between the fourth switching device and the third switching device.
[0363] 607b-608b. Similar to actions of the third switching device in
[0364] i is a natural number, and i is greater than or equal to 4.
[0365] When the second service device is a destination service device, step 610-611 is executed.
[0366] When the second service device is not a destination service device, steps 610 to 615 are executed.
[0367] 610-611. The second switching device deletes the identifier of the second service device and sends the data processed by the first service device to the second service device, according to the third forwarding rule.
[0368] 612-613. The second switching device receives the data processed by the second service device, and sends the data processed by the second service device and an identifier of a third service device to a next-hop switching device of the second service device according to a fifth forwarding rule (for example, a reactive flow entry in Table 12) received from the SDN control device. The second switching device may further send path information to the next-hop switching device according to the fifth forwarding rule received from the SDN control device.
[0369] An action of the second switching device is similar to the action of the first switching device in
[0370] 614-615. It is assumed that there are j service devices, and a j.sup.th service device is connected to a j.sup.th switching device, where j is a natural number, and j is greater than or equal to 3.
[0371] The j.sup.th switching device receives the data processed by a previous-hop (j−1.sup.th) service device and an identifier of the j.sup.th service device; and according to an indication of a forwarding rule of the SDN control device, deletes the identifier of the j.sup.th service device and sends the data processed by the j−1.sup.th service device to the j.sup.th service device.
[0372] An action of the j.sup.th switching device is similar to that of the first switching device. Details are not described again.
[0373] Step 614-615 is repeatedly executed according to a network topology status until the data is transmitted to a last-hop service device (that is, a destination service device).
[0374] An example of one SDN control device is used for description in the foregoing embodiment. The present invention is also applicable to a cluster that includes multiple SDN control devices.
[0375] In a scenario in which a service needs to be implemented, the method provided in this embodiment of the present invention can also achieve an effect of improving a resource utilization rate of a device in SDN. After receiving a first forwarding rule (which may also be referred to as a proactive flow entry), a first switching device does not need to acquire a first forwarding rule each time data is being forwarded. Accordingly, an SDN control device does not need to deliver a first forwarding rule to the first switching device each time the data is being forwarded, that is, the first forwarding rule may be reused during subsequent data forwarding. This reduces a quantity of first forwarding rules delivered by the SDN control device to the first switching device, and accordingly reduces a device resource occupation rate of the SDN control device; and network bandwidth between the SDN control device and the first switching device is also reduced accordingly. In addition, the first switching device does not need to receive a first forwarding rule each time the data is being forwarded, which improves a resource utilization rate of the first switching device.
[0376] In addition, when different service devices need to be passed through during data forwarding, data forwarding is implemented according to order of all the service devices on a service chain. During a data forwarding process, if a switching device (for example, the aggregation switching device 1, the core switching device, or the aggregation switching device 2 in
[0377] Referring to
[0378] The network communications unit 801 is configured to: receive a request for a forwarding rule from the first switching device, where the request carries characteristic information of the data; and send the request to the forwarding rule management unit 802.
[0379] The forwarding rule management unit 802 is configured to determine a first forwarding rule and an n.sup.th forwarding rule according to the characteristic information of the data, where the first forwarding rule is used to instruct the first switching device to send, to a second switching device connected to the first switching device, the data and information about a path from the first switching device to the n.sup.th switching device according to the first forwarding rule, and the n.sup.th forwarding rule is used to instruct the n.sup.th switching device to forward the data to the destination end.
[0380] The network communications unit 801 is further configured to: return the first forwarding rule to the first switching device, and return the n.sup.th forwarding rule to the n.sup.th switching device.
[0381] The SDN control device 800 may further include a topology management unit 803, which is configured to: collect network topology information of the SDN system, and send the network topology information to the forwarding rule management unit 802.
[0382] The forwarding rule management unit 802 is further configured to determine a second forwarding rule according to the network topology information, where the second forwarding rule is used to instruct the second switching device to: obtain a port, on the second switching device, of a third switching device by using a signpost of the third switching device in the information about the path from the first switching device to the n.sup.th switching device, and forward the data to the third switching device through the port, on the second switching device, of the third switching device, where the third switching device is a switching device connected to the second switching device. A match condition of the second forwarding rule is the signpost of the third switching device. An execution action of the second forwarding rule is to forward the data to the third switching device through the port, on the second switching device, of the third switching device. The port, on the second switching device, of the third switching device is obtained by using the signpost of the third switching device. The network communications unit 801 is further configured to deliver the second forwarding rule to the second switching device.
[0383] The forwarding rule management unit 802 is further configured to determine an i.sup.th forwarding rule according to the network topology information of the SDN system, where the i.sup.th forwarding rule is used to instruct an i.sup.th switching device to: obtain a port, on the i.sup.th switching device, of an i+1.sup.th switching device by using a signpost of the i+1.sup.th switching device in information about a path from an i−1.sup.th switching device to the n.sup.th switching device, and forward the data to the i+1.sup.th switching device through the port, on the i.sup.th switching device, of the i+1.sup.th switching device. The information about the path from the i−1.sup.th switching device to the n.sup.th switching device includes signposts of the i+1.sup.th switching device to the n.sup.th switching device that are on a data forwarding path. If the i+1.sup.th switching device is not the n.sup.th switching device, the i.sup.th forwarding rule is further used to instruct the i.sup.th switching device to: remove the signpost of the i+1.sup.th switching device, and send, to the i+1.sup.th switching device, information about a path from the i.sup.th switching device to the n.sup.th switching device, where the information about the path from the i.sup.th switching device to the n.sup.th switching device includes signposts of an i+2.sup.th switching device to the n.sup.th switching device that are on the data forwarding path, the i+2.sup.th switching device is a switching device connected to the i+1.sup.th switching device, i is an integer greater than or equal to 3, and i is less than n. The network communications unit 801 is further configured to deliver the i.sup.th forwarding rule to the i.sup.th switching device.
[0384] The forwarding rule management unit 802 is specifically configured to select the data forwarding path according to the characteristic information of the data and the network topology information of the SDN system, where the information about the path from the first switching device to the n.sup.th switching device includes signposts of the third switching device to the n.sup.th switching device that are on the data forwarding path.
[0385] The signpost of the third switching device includes at least one of the following: a device identifier of the third switching device, an identifier of the port, on the second switching device, of the third switching device, or an identifier of a link between the third switching device and the second switching device.
[0386] For a function of the SDN control device 800, specifically refer to related descriptions of the SDN control device in the data forwarding method 1. Details are not described herein again.
[0387] Because a second switching device and an i.sup.th switching device (backbone switching devices) receive proactive flow entries, and there is no need to deliver a flow entry to a backbone switching device each time data is being forwarded, that is, the proactive flow entries may be reused during subsequent data forwarding, a quantity of flow entries delivered to the backbone switching device is reduced, a device resource occupation rate of an SDN control device is reduced accordingly, and network bandwidth between the SDN control device and the backbone switching device is also reduced accordingly.
[0388] Referring to
[0389] The topology management unit 903 is configured to: collect network topology information of the SDN system, and send the network topology information to the forwarding rule management unit 902.
[0390] The forwarding rule management unit 902 is configured to determine a first forwarding rule according to the network topology information of the SDN system, where the first forwarding rule is used to instruct the first switching device to delete an identifier of the first service device and send data to the first service device for processing.
[0391] The network communications unit 901 is configured to: receive a request for a forwarding rule, where the request carries characteristic information; and send the request to the forwarding rule management unit 902.
[0392] The forwarding rule management unit 902 is further configured to determine a second forwarding rule according to the characteristic information carried in the request, where when the first switching device is connected to the second switching device, the second forwarding rule is used to instruct the first switching device to send the data processed by the first service device and an identifier of the second service device to the second switching device; or when the first switching device is not connected to the second switching device, the second forwarding rule is used to instruct the first switching device to send, to a third switching device, the data processed by the first service device, an identifier of the second service device, and information about a path from the first switching device to the second switching device, where the third switching device is a switching device connected to the first switching device.
[0393] The network communications unit 901 is further configured to return the first forwarding rule and the second forwarding rule to the first switching device.
[0394] The forwarding rule management unit 902 is further configured to determine a third forwarding rule according to the network topology information of the SDN system, where when the second service device is a destination service device, the third forwarding rule is used to instruct the second switching device to: delete the identifier of the second service device and forward the data processed by the first service device to the second service device, according to the third forwarding rule; or when the second service device is not a destination service device, the third forwarding rule is used to instruct the second switching device to: delete the identifier of the second service device and forward the data processed by the first service device to the second service device, according to the third forwarding rule; and receive the data processed by the second service device. The network communications unit 901 is further configured to return the third forwarding rule to the second switching device.
[0395] The forwarding rule management unit 902 is further configured to determine a fourth forwarding rule according to the network topology information of the SDN system, where the fourth forwarding rule is used to instruct the third switching device to: determine a port, on the third switching device, of a fourth switching device according to a signpost of the fourth switching device, and forward the data processed by the first service device to the fourth switching device through the port, on the third switching device, of the fourth switching device.
[0396] If the fourth switching device is not the second switching device, the fourth forwarding rule is further used to instruct the third switching device to send, to the fourth switching device, information about a path from the third switching device to the second switching device. The fourth forwarding rule specifically instructs the third switching device to: remove the signpost of the fourth switching device from the information about the path from the first switching device to the second switching device to obtain the information about the path from the third switching device to the second switching device, and send, to the fourth switching device, the information about the path from the third switching device to the second switching device.
[0397] The network communications unit 901 is further configured to return the fourth forwarding rule to the third switching device.
[0398] For a function of the SDN control device 900, specifically refer to related descriptions of the SDN control device in the data forwarding method 2. Details are not described herein again.
[0399] After receiving a first forwarding rule (which may also be referred to as a proactive flow entry), a first switching device does not need to acquire a first forwarding rule each time data is being forwarded. Accordingly, an SDN control device does not need to deliver a first forwarding rule to the first switching device each time the data is being forwarded, that is, the first forwarding rule may be reused during subsequent data forwarding. This reduces a quantity of first forwarding rules delivered by the SDN control device to the first switching device, and accordingly reduces a device resource occupation rate of the SDN control device; and network bandwidth between the SDN control device and the first switching device is also reduced accordingly. In addition, the first switching device does not need to receive a first forwarding rule each time the data is being forwarded, which improves a resource utilization rate of the first switching device.
[0400] In addition, when different service devices need to be passed through during data forwarding, data forwarding is implemented according to order of all the service devices on a service chain. During a data forwarding process, the SDN control device delivers a proactive flow entry to a switching device (for example, the aggregation switching device 1, the core switching device, or the aggregation switching device 2 in
[0401] SDN System 3
[0402] In the foregoing data forwarding method 1, there is one source end and one destination end, and the method may be referred to as a unicast solution. A data forwarding method in which one source end and multiple destination end exist may be referred to as a multicast or broadcast solution. Multicast is used as an example for detailed description in the following.
[0403] Multiple members may form a multicast group. Data (or a service packet) sent by a member in a multicast group can be received only by a member in the multicast group. A member in a multicast group may be a device such as a mobile terminal (which includes but is not limited to a laptop, a mobile phone, a tablet, a palmtop computer, a POS, an MP3, and a navigator), a fixed terminal (which includes but is not limited to a desktop computer, a printer, a fax machine, a digital projector, and a digital television), or a wireless access terminal.
[0404] A manner of implementing the data forwarding method in a multicast group according to this embodiment of the present invention is described by using an example in which the multicast group has four members, where the four members are connected to different switching devices and implement data forwarding by using a backbone switching device in a DC network. As shown in
[0405] In
[0406] The multicast group member 1 sends multicast packets to other members, where: a multicast packet sent to the member 2 passes through the virtual switching device 1 from a port 2, goes out from a port 5, passes through the access switching device 1 from a port 7, and goes out from a port 1; a multicast packet sent to the member 3 passes through the virtual switching device 1 from the port 2, goes out from the port 5, passes through the access switching device 1 from the port 7, goes out from a port 4, passes through an aggregation switching device 1 from a port 6, goes out from a port 3, passes through a core switching device from a port 5, goes out from a port 9, passes through an aggregation switching device 2 from a port 8, goes out from a port 4, passes through the access switching device 3 from a port 5, and goes out from a port 8; a multicast packet sent to the member 4 passes through the virtual switching device 1 from the port 2, goes out from the port 5, passes through the access switching device 1 from the port 7, goes out from the port 4, passes through the aggregation switching device 1 from the port 6, goes out from the port 3, passes through the core switching device from the port 5, goes out from the port 9, passes through the aggregation switching device 2 from the port 8, goes out from a port 7, passes through an access switching device 2 from a port 6, goes out from a port 8, passes through the virtual switching device 2 from a port 3, and goes out from a port 1.
[0407] The multicast group member 2 sends multicast packets to other members, where: a multicast packet sent to the member 1 passes through the access switching device 1 from the port 1, goes out from the port 7, passes through the virtual switching device 1 from the port 5, and goes out from the port 2; a multicast packet sent to the member 3 passes through the access switching device 1 from the port 1, goes out from the port 4, passes through the aggregation switching device 1 from the port 6, goes out from the port 3, passes through the core switching device from the port 5, goes out from the port 9, passes through the aggregation switching device 2 from the port 8, goes out from the port 4, passes through the access switching device 3 from the port 5, and goes out from the port 8; a multicast packet sent to the member 4 passes through the access switching device 1 from the port 1, goes out from the port 4, passes through the aggregation switching device 1 from the port 6, goes out from the port 3, passes through the core switching device from the port 5, goes out from the port 9, passes through the aggregation switching device 2 from the port 8, goes out from the port 7, passes through the access switching device 2 from the port 6, goes out from the port 8, passes through the virtual switching device 2 from the port 3, and goes out from the port 1.
[0408] The multicast group member 3 sends multicast packets to other members, where: a multicast packet sent to the member 1 passes through the access switching device 3 from the port 8, goes out from the port 5, passes through the aggregation switching device 2 from the port 4, goes out from the port 8, passes through the core switching device from the port 9, goes out from the port 5, passes through the aggregation switching device 1 from the port 3, goes out from the port 6, passes through the access switching device 1 from the port 4, goes out from the port 7, passes through the virtual switching device 1 from the port 5, and goes out from the port 2; a multicast packet sent to the member 2 passes through the access switching device 3 from the port 8, goes out from the port 5, passes through the aggregation switching device 2 from the port 4, goes out from the port 8, passes through the core switching device from the port 9, goes out from the port 5, passes through the aggregation switching device 1 from the port 3, goes out from the port 6, passes through the access switching device 1 from the port 4, and goes out from the port 1; a multicast packet sent to the member 4 passes through the access switching device 3 from the port 8, goes out from the port 5, passes through the aggregation switching device 2 from the port 4, goes out from the port 7, passes through the access switching device 2 from the port 6, goes out from the port 8, passes through the virtual switching device 2 from the port 3, and goes out from the port 1.
[0409] The multicast group member 4 sends multicast packets to other members, where: a multicast packet sent to the member 1 passes through the virtual switching device 2 from the port 1, goes out from the port 3, passes through the access switching device 2 from the port 8, goes out from the port 6, passes through the aggregation switching device 2 from the port 7, goes out from the port 8, passes through the core switching device from the port 9, goes out from the port 5, passes through the aggregation switching device 1 from the port 3, goes out from the port 6, passes through the access switching device 1 from the port 4, goes out from the port 7, passes through the virtual switching device 1 from the port 5, and goes out from the port 2; a multicast packet sent to the member 2 passes through the virtual switching device 2 from the port 1, goes out from the port 3, passes through the access switching device 2 from the port 8, goes out from the port 6, passes through the aggregation switching device 2 from the port 7, goes out from the port 8, passes through the core switching device from the port 9, goes out from the port 5, passes through the aggregation switching device 1 from the port 3, goes out from the port 6, passes through the access switching device 1 from the port 4, and goes out from the port 1; a multicast packet sent to the member 3 passes through the virtual switching device 2 from the port 1, goes out from the port 3, passes through the access switching device 2 from the port 8, goes out from the port 6, passes through the aggregation switching device 2 from the port 7, goes out from the port 4, passes through the access switching device 3 from the port 5, and goes out from the port 8.
[0410] Data Forwarding Method 3
[0411] 1. Proactive Flow Entry
[0412] To implement data forwarding in a multicast manner, an SDN control device needs to collect or update multicast group information, which is similar to collecting or updating network topology information. The following describes three cases:
[0413] When a new member is to join a multicast group, a process of triggering updating of the multicast group information is as follows: The new multicast group member sends, to a multicast router, a request for joining the multicast group; because a switching device that receives the request for joining the multicast group has no flow entry delivered by the SDN control device, the switching device reports a PacketIn packet to the SDN control device; the SDN control device simulates the multicast router, analyzes the request for joining the multicast group from the multicast group member, and then updates the multicast group information; the SDN control device sends, to the switching device, a response to the joining of the multicast group; and the switching device that reports the PacketIn packet sends, to the member that is to join the multicast group, the response to the joining of the multicast group.
[0414] When a multicast group member is to exit from a multicast group, a process of triggering updating of the multicast group information is as follows: The member that needs to exit from the multicast group sends, to a multicast router, a request for exiting from the multicast group; because a switching device that receives the request for exiting from the multicast group has no matched flow entry, the switching device reports a PacketIn packet to the SDN control device; the SDN control device simulates the multicast router, analyzes the request, from the multicast group member, for exiting from the multicast group, and then updates the multicast group information; the SDN control device sends, to the switching device, a response to the exiting from the multicast group; and the switching device that reports the
[0415] PacketIn packet sends, to the member that is to exit from the multicast group, the response to the exiting from the multicast group.
[0416] When the SDN control device queries a multicast group, a process of triggering updating of the multicast group information is as follows: The SDN control device periodically sends a PacketOut multicast group query request to a switching device connected to a multicast group member; the switching device sends the multicast group query request to a port on which the multicast group member is located; the multicast group member returns a response to multicast group query; because the switching device has no matched flow entry, the switching device reports a PacketIn packet to the SDN control device; and the SDN control device analyzes a multicast group query result, and updates the multicast group information.
[0417] After learning the multicast group information, the SDN control device needs to create a multicast group identifier. The multicast group identifier in this embodiment of the present invention is unique, and is used to distinguish from another multicast group in a DC network or another type of object that can have a device identifier; the multicast group identifier may be independent, or may be combined with another device identifier, and determines a packet forwarding path. For example, in this embodiment of the present invention, the multicast group identifier created by the SDN control device for the multicast group is 4.
[0418] Similar to the proactive flow entries in the data forwarding method 1, the SDN control device needs to deliver proactive flow entries to all backbone switching devices according to network topology information. For details, refer to the description in the data forwarding method 1, and details are not described again.
[0419] As shown in
[0420] When a multicast group member changes, updating of the switching device on which the rendezvous point is located is triggered, for example, adding or deleting information about the switching device on which the rendezvous point in the multicast group is located. The SDN control device updates information about the rendezvous point in the multicast group. The information about the rendezvous point in the multicast group includes multicast group members, and corresponding multicast packet ingress ports and egress ports. The multicast packet is sent from the multicast source to the switching device on which the rendezvous point in the multicast group is located, where a receive port, on the switching device on which the rendezvous point in the multicast group is located, of the multicast packet is an ingress port. The multicast packet is sent from the switching device on which the rendezvous point in the multicast group is located to a multicast receiver, where a transmit port, on the switching device on which the rendezvous point in the multicast group is located, of the multicast packet is an egress port.
[0421] After acquiring the switching device on which the rendezvous point is located, the SDN control device generates information about the rendezvous point in the multicast group.
TABLE-US-00013 TABLE 13 Switching devices on Multicast which rendezvous group Egress Ingress points are located members ports ports Access switching Member 1 7 4 device 1 Member 2 1 4 Members 3 and 4 4 1, 7 Aggregation Member 3 4 8 switching Member 4 7 8 device 2 Members 1 and 2 8 4, 7
[0422] In Table 13, the access switching device 1 disposes the egress port 7 and the ingress port 4 for the member 1, disposes the egress port 1 and the ingress port 4 for the member 2, and disposes the egress port 4 and the ingress ports 1 and 7 for the members 3 and 4. The aggregation switching device 2 disposes the egress port 4 and the ingress port 8 for the member 3, disposes the egress port 7 and the ingress port 8 for the member 4, and disposes the egress port 8 and the ingress ports 4 and 7 for the members 1 and 2.
[0423] For the switching device on which the rendezvous point is located, a proactive flow entry delivered by the SDN control device needs to meet the following characteristics:
[0424] a match condition: an ingress port of the rendezvous point in the multicast group and a multicast group identifier; and
[0425] execution actions:
[0426] sending a multicast packet to a receiver: copying the multicast packet, removing the multicast group identifier from the copied multicast packet, and sending an original multicast packet to an egress port of the receiver; and sending the multicast packet to a next rendezvous point in the multicast group: copying the multicast packet, encapsulating a forwarding path to the next rendezvous point in the multicast group for the copied multicast packet, and sending the encapsulated multicast packet to a next-hop egress port.
[0427] For example, proactive flow entries delivered to the switching devices on which the rendezvous points are located in
TABLE-US-00014 TABLE 14 Switching devices Proactive flow entries Access Match condition: port 4 and multicast group identifier switching Copy the multicast packet and separately execute actions: device 1 (1). removing the multicast group identifier and sending an original multicast packet to the port 1; (2) sending an original multicast packet to the port 7 Match condition: port 7 and multicast group identifier Copy the multicast packet and separately execute actions: (1). removing the multicast group identifier and sending an original multicast packet to the port 1; (2). encapsulating a forwarding path to the aggregation switching device 2 and sending, to the port 4, the multicast packet in which the forwarding path is encapsulated Aggregation Match condition: port 4 and multicast group identifier switching Copy the multicast packet and separately execute actions: device 2 (1). encapsulating a forwarding path to the virtual switching device 2 and sending, to the port 7, the multicast packet in which the forwarding path is encapsulated; (2). encapsulating a forwarding path to the access switching device 1 and sending, to the port 8, the multicast packet in which the forwarding path is encapsulated Match condition: port 7 and multicast group identifier Copy the multicast packet and separately execute actions: (1) sending an original multicast packet to the port 4; (2). encapsulating a forwarding path to the access switching device 1 and sending, to the port 8, the multicast packet in which the forwarding path is encapsulated Match condition: port 8 and multicast group identifier Copy the multicast packet and separately execute actions: (1) sending an original multicast packet to the port 4; (2). encapsulating a forwarding path to the virtual switching device 2 and sending, to the port 7, the multicast packet in which the forwarding path is encapsulated
[0428] In Table 14, the SDN control device copies and forwards multicast packets received by the access switching device 1 from different ports. Multicast packets sent by the members 3 and 4 are received on the port 4, where: a multicast packet that goes to the member 1 is sent to the port 7; and for a multicast packet that goes to the member 2, an outer signpost is removed, and the multicast packet from which the signpost is removed is sent to the port 1. Multicast packets sent by the member 1 are received on the port 7, where: for a multicast packet that goes to the member 2, an outer signpost is removed, and the multicast packet from which the signpost is removed is sent to the port 1; and for multicast packets that go to the members 3 and 4, the forwarding path to a switching device (the aggregation switching device 2) on which a rendezvous point is located is encapsulated, and the multicast packets in which the forwarding path is encapsulated are sent to the port 4. The SDN control device copies and forwards multicast packets received by the aggregation switching device 2 from different ports. Multicast packets sent by the member 3 are received on the port 4, where: for a multicast packet that goes to the member 4, the forwarding path to the virtual switching device 2 is encapsulated, and the multicast packet in which the forwarding path is encapsulated is sent to the port 7; and for multicast packets that go to the members 1 and 2, the forwarding path to the access switching device 1 is encapsulated, and the multicast packets in which the forwarding path is encapsulated are sent to the port 8. Multicast packets sent by the member 4 are received on the port 7, where: a multicast packet that goes to the member 3 is sent to the port 4; and for multicast packets that go to the members 1 and 2, the forwarding path to the access switching device 1 on which the rendezvous point in the multicast group is located is encapsulated, and the multicast packets in which the forwarding path is encapsulated are sent to the port 8. Multicast packets sent by the members 1 and 2 are received on the port 8, where: a multicast packet that goes to the member 3 is sent to the port 4; and for a multicast packet that goes to the member 4, the forwarding path to the virtual switching device 2 is encapsulated, and the multicast packet in which the forwarding path is encapsulated is sent to the port 7.
[0429] Another proactive flow entry delivered by the SDN control device to the switching device connected to the multicast destination end (which is also referred to as a receiver) needs to meet the following characteristics:
[0430] a match condition: a multicast packet port and a multicast group identifier; and
[0431] an execution action: removing the multicast group identifier from a multicast packet and sending an original multicast packet to the destination end.
[0432] For example, flow entries delivered to switching devices connected to multicast destination ends (which are also referred to as receivers) in
TABLE-US-00015 TABLE 15 Devices devices Proactive flow entries Virtual Match condition: port 5 and multicast group identifier switching Execution action: removing the multicast group device 1 identifier and sending an original multicast packet to the port 2 Access Match condition: port 5 and multicast group identifier switching Execution action: removing the multicast group device 3 identifier and sending an original multicast packet to the port 8 Virtual Match condition: port 3 and multicast group identifier switching Execution action: removing the multicast group device 2 identifier and sending an original multicast packet to the port 1
[0433] In Table 15, for a multicast packet sent to the member 1, the SDN control device delivers a proactive flow entry to the virtual switching device 1 to match the multicast packet port 5 and the multicast group 1 identifier, remove the multicast group 1 identifier from the multicast packet, and send the original multicast packet to the port 2. For a multicast packet sent to the member 3, the SDN control device delivers a proactive flow entry to the access switching device 3 to match the multicast packet port 5 and the multicast group 1 identifier, remove the multicast group 1 identifier from the multicast packet, and send the original multicast packet to the port 8. For a multicast packet sent to the member 4, the SDN control device customizes a proactive flow entry for the virtual switching device 2 to match the multicast packet port 3 and the multicast group 1 identifier, remove the multicast group 1 identifier, and send the original multicast packet to the port 1.
[0434] Similar to implementation of the data forwarding method 1, when a port status of a switching device changes, the SDN control device updates the multicast group information. As shown in
[0435] When an egress port is added to the switching device on which the rendezvous point is located, a multicast proactive flow entry that uses the egress port as a transmit port and uses another ingress port as a receive port is added to the switching device on which the rendezvous point is located.
[0436] When an egress port is deleted from the switching device on which the rendezvous point is located, all multicast proactive flow entries that use the egress port as a transmit port are deleted from the switching device on which the rendezvous point is located.
[0437] When an ingress port is added to the switching device on which the rendezvous point is located, a multicast proactive flow entry that uses another egress port as a transmit port and uses the ingress port as a receive port is added to the switching device on which the rendezvous point is located.
[0438] When an egress port is deleted from the switching device on which the rendezvous point is located, all multicast proactive flow entries that use the egress port as a receive port are deleted from the switching device on which the rendezvous point is located.
[0439] 2. Reactive Flow Entry
[0440] When sending data or a service packet to another member in a multicast group, a member in the multicast group sends a multicast packet to a switching device connected to the member. Because the connected switching device has no matched flow entry, the switching device reports a PacketIn packet to the SDN control device. The SDN control device analyzes a characteristic of the multicast packet, and selects a proper forwarding path according to current network topology information and multicast group information. A reactive flow entry meets the following characteristics:
[0441] a match condition: a port and a packet characteristic; and
[0442] an execution action: successively encapsulating, for a multicast packet, a multicast group identifier and a forwarding path to a switching device on which a next rendezvous point in the multicast group is located, and sending the encapsulated multicast packet to a next-hop egress port.
[0443] For example, in the network architecture diagram shown in
TABLE-US-00016 TABLE 16 Switching devices Reactive flow entries Virtual Match condition: port 2 and multicast packet characteristic switching Execution action: encapsulating the multicast group device 1 identifier and sending the encapsulated multicast packet to the port 5 Access Match condition: port 1 and multicast packet characteristic switching Encapsulate the multicast group identifier, copy the device 1 multicast packet, and separately execute actions: (1). encapsulating a forwarding path to the aggregation switching device 2 and sending the encapsulated multicast packet to the port 4; (2) sending the encapsulated multicast packet to the port 7 Access Match condition: port 8 and multicast packet characteristic switching Execution action: encapsulating the multicast group device 3 identifier and sending the encapsulated multicast packet to the port 5 Virtual Match condition: port 1 and multicast packet characteristic switching Execution action: encapsulating the multicast group device 2 identifier and a forwarding path to the aggregation switching device 2, and sending the encapsulated multicast packet to the port 3
[0444] In Table 16, for a multicast packet sent by the member 1, the reactive flow entry instructs the virtual switching device 1 to match the packet port 2 and the packet characteristic, encapsulate the multicast group 1 identifier into the original multicast packet, and send the encapsulated multicast packet to the port 5. For a multicast packet sent by the member 2 to the member 1, the reactive flow entry instructs the access switching device 1 to match the packet port 1 and the packet characteristic, encapsulate the multicast group 1 identifier into the original multicast packet, and send the encapsulated multicast packet to the port 7. For multicast packets sent by the member 2 to the members 3 and 4, the reactive flow entry instructs the access switching device 1 to match the packet port 1 and the packet characteristic, successively encapsulate the multicast group 1 identifier and the forwarding path to the aggregation switching device 2 into the original multicast packet, and send the encapsulated multicast packet to the port 4. For a multicast packet sent by the member 3, the reactive flow entry instructs the access switching device 3 to match the packet port 8 and the packet characteristic, encapsulate the multicast group 1 identifier into the original multicast packet, and send the encapsulated multicast packet to the port 5. For a multicast packet sent by the member 4, the reactive flow entry instructs the virtual switching device 2 to match the packet port 1 and the packet characteristic, successively encapsulate the multicast group 1 identifier and the forwarding path to the aggregation switching device 2 into the original multicast packet, and send the encapsulated multicast packet to the port 3. In actual implementation, it is possible that there is only one member used as a source end to send a packet; therefore, only one reactive flow entry in the foregoing Table 16 is used.
[0445] 3. Data Forwarding
[0446] That the switching device on which the rendezvous point in the multicast group is located is represented as a first switching device (the access switching device 1 or the aggregation switching device 2 in
[0447] The first switching device is configured to receive the data and a multicast group identifier.
[0448] The first switching device is configured to execute one of the following according to a forwarding rule (for example, a proactive flow entry in Table 14) received from the SDN control device in the SDN system:
[0449] when the first switching device is connected to the destination end, the first switching device is configured to: delete the multicast group identifier, and forward the data to the destination end;
[0450] when the first switching device is connected to a second switching device, the first switching device is configured to forward the data and the multicast group identifier to the second switching device, where the second switching device is a switching device connected to the destination end or a switching device on which another rendezvous point in the multicast group is located; or
[0451] when the first switching device is not connected to the destination end or a second switching device, the first switching device is configured to send, to a third switching device, the data, the multicast group identifier, and information about a path from the first switching device to the second switching device, where the second switching device is a switching device connected to the destination end or a switching device on which another rendezvous point in the multicast group is located, and the third switching device is a switching device connected to the first switching device.
[0452] The information about the path from the first switching device to the second switching device includes signposts of a fourth switching device to the second switching device that are on the data forwarding path, where the fourth switching device is a switching device, on the data forwarding path, connected to the third switching device.
[0453] The first switching device is configured to successively push the multicast group identifier and the signposts of the second switching device to the fourth switching device that are on the data forwarding path into a protocol stack, where the multicast group identifier is shifted into the bottom of the stack, a signpost of the second switching device is shifted into an upper layer of the bottom of the stack, and a signpost of the fourth switching device is shifted into the top of the stack, so that the protocol stack carries the path information and the multicast group identifier. Encapsulation of the multicast group identifier is similar to that of an identifier of a service device, and details are not described again.
[0454] The third switching device is configured to identify, according to the signpost of the fourth switching device, that a next-hop switching device of the data is the fourth switching device, where the signpost of the fourth switching device includes at least one of the following: a device identifier of the fourth switching device, an identifier of an egress port, on the third switching device, of the fourth switching device, or an identifier of a link between the fourth switching device and the third switching device.
[0455] The first switching device is further configured to copy the data according to the forwarding rule, where when the first switching device is neither connected to the another destination end nor a fifth switching device, the first switching device sends, to a sixth switching device, the data, the multicast group identifier, and information about a path from the first switching device to the fifth switching device, where the sixth switching device is connected to the first switching device, and the fifth switching device is a switching device connected to the another destination end or a switching device on which another rendezvous point is located.
[0456] In addition, it should be noted that, similar to actions of the third switching device in
[0457] A multicast or broadcast solution is similar to the data forwarding method 2. Because the data can arrive at multiple destination ends, when delivering a forwarding path, the SDN control device cannot deliver forwarding paths to all the destination ends at a time, but needs to divide a path into multiple segments according to a switching device on which each rendezvous point is located. A start of each segment is a switching device on which the source end is located or a switching device on which a rendezvous point is located, an end is a switching device on which the destination end is located or a switching device on which another rendezvous point is located, and a path delivery manner of each segment is similar to an implementation manner shown in the data forwarding method 1. Therefore, for a specific data forwarding procedure, refer to the descriptions in the data forwarding method 1 and the data forwarding method 2, and details are not described again.
[0458] In a scenario in which multicast needs to be implemented, the method provided in this embodiment of the present invention can also reach an effect of improving a resource utilization rate of a device in SDN. After receiving a forwarding rule (which may also be referred to as a flow entry), a switching device (for example, a first switching device) on which a rendezvous point is located does not need to acquire a forwarding rule each time data is being forwarded. Accordingly, an SDN control device does not need to deliver a forwarding rule to the first switching device each time the data is being forwarded, that is, the forwarding rule may be reused during subsequent data forwarding. This reduces a quantity of forwarding rules delivered by the SDN control device to the first switching device, and accordingly reduces a device resource occupation rate of the SDN control device; and network bandwidth between the SDN control device and the first switching device is also reduced accordingly. In addition, the first switching device does not need to receive a forwarding rule each time the data is being forwarded, which improves a resource utilization rate of the first switching device.
[0459] In addition, during a data forwarding process, an aggregation switching device (for example, the access switching device 1 and the aggregation switching device 2 in
[0460] Network Security
[0461] A DC generally supports multiple tenants, and allows the tenants to share a physical network of the DC and construct respective independent logical networks. The logical networks of the tenants are mutually separated, and cyberspaces may overlap. Therefore, to improve network security, the SDN control device may further allocate a tenant identifier to a tenant. Then, in step 204, the SDN control device further sends the tenant identifier to the first switching device by delivering the first forwarding rule.
[0462] In the embodiment shown in
Table 6′
[0463]
TABLE-US-00017 TABLE 6′ Switching device Reactive flow entry Virtual switching Match condition: characteristic information of the data device 1 Execution action: encapsulating the tenant identifier and information about a path from the virtual switching device 1 to the virtual switching device 2 and sending, to the port 5, the data in which the tenant identifier and the path information are encapsulated
[0464] The reactive flow entry 1 includes the following information:
[0465] a match condition: the characteristic information of the data; and
[0466] an execution action: encapsulating the tenant identifier and the information about the path from the virtual switching device 1 to the virtual switching device 2 and sending, to the port 5, the data in which the tenant identifier and the path information are encapsulated. The virtual switching device 2 is a switching device connected to the destination end of the data.
[0467] In step 205, the SDN control device further delivers, to the n.sup.th switching device by delivering the n.sup.th forwarding rule, a match condition for matching the tenant identifier and an execution action of removing the tenant identifier. Table 7 changes into Table 7′.
TABLE-US-00018 TABLE 7 Switching device Reactive flow entry Virtual Match condition: network identifier and characteristic switching information of the data device 2 Execution action: removing the network identifier and sending, to the port 1, the data from which the network identifier is removed
[0468] In step 206, the first switching device shifts the tenant identifier into the bottom of the stack, and shifts the signpost of the n.sup.th switching device into an upper layer of the bottom of the stack. The first switching device further sends, according to the first forwarding rule, the tenant identifier to the second switching device connected to the first switching device, so as to send the tenant identifier to the n.sup.th switching device by using the second switching device.
[0469] An action of a backbone switching device is similar to steps 207 to 208, and details are not described again.
[0470] In step 209, the n.sup.th switching device determines the destination end of the data according to the n.sup.th forwarding rule by using the tenant identifier and the characteristic information of the data, and then sends the data to the destination end.
[0471] To implement network security, it is only needed to change the reactive flow entry delivered by the SDN control device to the edge switching device, and the SDN control device does not need to deliver a proactive flow entry to the backbone switching device; therefore, it is easy to implement. In addition, according to the foregoing method for implementing data forwarding, after receiving a forwarding rule (for example, a proactive flow entry in the foregoing embodiment), a backbone switching device does not need to acquire a forwarding rule each time data is being forwarded, which reduces resource overheads of a control device and a switching device, and improves a utilization rate of resources in an SDN system.
[0472] Referring to
[0473] The processor 601 is configured to execute a program. The program in this embodiment may include program code, where the program code includes a computer operation instruction. The processor may be a central processing unit CPU, or one or more integrated circuits configured to implement this embodiment of the present invention. The program executed by the processor is a program corresponding to each step executed by the SDN control device in the foregoing embodiments.
[0474] The memory 602 is configured to store the program executed by the processor.
[0475] The communications port 603 is configured to communicate with a switching device.
[0476] For specific functions, refer to the descriptions of the SDN control device in the foregoing embodiments.
[0477] A person of ordinary skill in the art may be aware that, in combination with the examples described in the embodiments disclosed in this specification, units and algorithm steps may be implemented by electronic hardware, computer software, or a combination thereof. To clearly describe the interchangeability between the hardware and the software, the foregoing has generally described compositions and steps of each example according to functions. Whether the functions are performed by hardware or software depends on particular applications and design constraint conditions of the technical solutions. A person skilled in the art may use different methods to implement the described functions for each particular application, but it should not be considered that the implementation goes beyond the scope of the present invention.
[0478] It may be clearly understood by a person skilled in the art that, for the purpose of convenient and brief description, for a detailed working process of the foregoing system, device, and unit, reference may be made to a corresponding process in the foregoing method embodiments, and details are not described herein again.
[0479] In the several embodiments provided in the present application, it should be understood that the disclosed system, device, and method may be implemented in other manners. For example, the described device embodiment is merely exemplary.
[0480] For example, the unit division is merely logical function division and may be other division in actual implementation. For example, a plurality of units or components may be combined or integrated into another system, or some features may be ignored or not performed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections may be implemented through some ports. The indirect couplings or communication connections between the devices or units may be implemented in electronic, mechanical, or other forms.
[0481] The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one position, or may be distributed on a plurality of network units. A part or all of the units may be selected according to actual needs to achieve the objectives of the solutions of the embodiments of the present invention.
[0482] In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each of the units may exist alone physically, or two or more units are integrated into one unit. The integrated unit may be implemented in a form of hardware, or may be implemented in a form of a software functional unit.
[0483] When the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, the integrated unit may be stored in a computer-readable storage medium. Based on such an understanding, the technical solutions of the present invention essentially, or the part contributing to the prior art, or all or a part of the technical solutions may be implemented in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for instructing a computer device (which may be a personal computer, a server, or a network device) to perform all or a part of the steps of the methods described in the embodiments of the present invention. The foregoing storage medium includes: any medium that can store program code, such as a USB flash drive, a removable hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk, or an optical disc.
[0484] The foregoing descriptions are merely specific embodiments of the present invention, but are not intended to limit the protection scope of the present invention. Any modification or replacement readily figured out by a person skilled in the art within the technical scope disclosed in the present invention shall fall within the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.