Methods and apparatuses are described for virtualizing routing of network traffic by offloading routing decisions to a controller in communication with a plurality of network devices. For load balancing applications, the controller may make up-front decisions as to both destination and route, rather than wait until traffic has been routed to a load balancing point before determining the destination.

Methods and apparatuses are described for virtualizing routing of network traffic by offloading routing decisions to a controller in communication with a plurality of network devices. For load balancing applications, the controller may make up-front decisions as to both destination and route, rather than wait until traffic has been routed to a load balancing point before determining the destination.

In a slice-based network, switches can be programmed to perform routing functions based on a slice identifier. The switch can receive a packet and determine a slice identifier for the packet based on packet header information. The switch can use the slice identifier to determine a next hop. Using the slice identifier with a multi-path table, the switch can select an egress interface for sending the packet to the next hop. The multi-path table can ensure that traffic for a slice stays on the same interface link to the next hop, even when a link aggregation group (“LAG”) is used for creation of a virtual channel across multiple interfaces or ports.

In a slice-based network, switches can be programmed to perform routing functions based on a slice identifier. The switch can receive a packet and determine a slice identifier for the packet based on packet header information. The switch can use the slice identifier to determine a next hop. Using the slice identifier with a multi-path table, the switch can select an egress interface for sending the packet to the next hop. The multi-path table can ensure that traffic for a slice stays on the same interface link to the next hop, even when a link aggregation group (“LAG”) is used for creation of a virtual channel across multiple interfaces or ports.

Techniques for load balancing communication sessions in a networked computing environment are described herein. The techniques may include establishing a first communication session between a client device and a first computing resource of a networked computing environment. Additionally, the techniques may include storing, in a data store, data indicating that the first communication session is associated with the first computing resource. The techniques may further include receiving, at a second computing resource of the networked computing environment, traffic associated with a second communication session that was sent by the client device, and based at least in part on accessing the data stored in the data store, establishing a traffic redirect such that the traffic and additional traffic associated with the second communication session is sent from the second computing resource to the first computing resource.

Techniques for load balancing communication sessions in a networked computing environment are described herein. The techniques may include establishing a first communication session between a client device and a first computing resource of a networked computing environment. Additionally, the techniques may include storing, in a data store, data indicating that the first communication session is associated with the first computing resource. The techniques may further include receiving, at a second computing resource of the networked computing environment, traffic associated with a second communication session that was sent by the client device, and based at least in part on accessing the data stored in the data store, establishing a traffic redirect such that the traffic and additional traffic associated with the second communication session is sent from the second computing resource to the first computing resource.

Described are programmable input output (IO) devices comprising: an match processing unit (MPU) and a memory unit. The MPU comprising at least one arithmetic logic unit (ALU). The memory unit having instructions stored thereon which, when executed by the respective programmable IO device, cause the programmable IO device to perform operations. These operations comprise: receiving, from an inbound interface, a packet comprising packet data for at least one range-based element; determining, via the MPU, a lookup result by performing a modified binary search on an interval binary search tree with the packet data to determine a longest prefix match (LPM), wherein the interval binary search tree maps the at least one range-based element to an associated data element; and classifying the packet based on the lookup result.

Technologies for dynamic accelerator selection include a compute sled. The compute sled includes a network interface controller to communicate with a remote accelerator of an accelerator sled over a network, where the network interface controller includes a local accelerator and a compute engine. The compute engine is to obtain network telemetry data indicative of a level of bandwidth saturation of the network. The compute engine is also to determine whether to accelerate a function managed by the compute sled. The compute engine is further to determine, in response to a determination to accelerate the function, whether to offload the function to the remote accelerator of the accelerator sled based on the telemetry data. Also the compute engine is to assign, in response a determination not to offload the function to the remote accelerator, the function to the local accelerator of the network interface controller.

Methods and systems are provided for performing lossy dropping and ECN marking in a flow-based network. The system can maintain state information of individual packet flows, which can be set up or released dynamically based on injected data. Each flow can be provided with a flow-specific input queue upon arriving at a switch. Packets of a respective flow are acknowledged after reaching the egress point of the network, and the acknowledgement packets are sent back to the ingress point of the flow along the same data path. As a result, each switch can obtain state information of each flow and perform per-flow packet dropping and ECN marking.

Some embodiments provide a network forwarding integrated circuit (IC) that includes at least one packet processing pipeline. The packet processing pipeline includes multiple match-action stages, at least one of which includes a stateful processing unit that operates at a line rate of the network forwarding IC. The stateful processing unit is configured to receive data stored in a memory location associated with a stateful table of the match-action stage. The data includes a set of values. The stateful processing unit is further configured to identify one of a maximum value and a minimum value from the set of values, and to output the identified value for use by a next match-action stage.