The invention is that of systems and methods for controllerless and distributed network connections to servers on a network by remote clients seeking their services. The invention comprises a central database where servers within a server group identified by a group name may post unique identifiers (UIDs) for retrieval by group clients configured with the group name, which may query the central database for server connection information such as uptime, downtime, and congestion in order to select a server for a preferential connection based on an overall availability profile as determined by the group client. In some embodiments, one server of the group is a “dummy server” that may aggregate and selectively transmit server information from other servers in the group, or other devices, and posts it to the central database for client access. The methods described herein eliminate a separate controller and thereby a single point of failure (SPOF).

Various example implementations are directed to circuits, apparatuses, and methods for providing virtual computing services. According to an example embodiment, an apparatus includes a computing server configured to provide a respective group of virtual servers for each of a plurality of accounts. Each of the accounts has a respective set of domain names and a respective settings file. The apparatus also includes a domain name server (DNS). The DNS is to dynamically map a respective set of domain names for each account to network addresses of the respective group of virtual servers, provided for the account. The DNS performs the mapping according to a mapping function indicated in the respective settings file of the account. The respective settings file of a first account accounts includes a mapping function that is different from a mapping function included in the respective settings file of a second account.

A method begins by a load balancing module of a distributed storage network (DSN) determining availability of a plurality of DSN processing units of a set of DSN processing units based on availability information associated with the plurality of DSN processing units and in response to determined availability, selecting a DSN processing unit form the set to process a data access request. The method continues with the load balancing module receiving an indication that the DSN processing unit is no longer available from the DSN processing unit while the DSN processing unit continues to process previously pending data access requests. The method continues with the load balancing module cancelling selection of the DSN processing unit to process the data access request; and receiving a second indication from the DSN processing unit indication that the DSN processing unit is available.

A method begins by a load balancing module of a distributed storage network (DSN) determining availability of a plurality of DSN processing units of a set of DSN processing units based on availability information associated with the plurality of DSN processing units and in response to determined availability, selecting a DSN processing unit form the set to process a data access request. The method continues with the load balancing module receiving an indication that the DSN processing unit is no longer available from the DSN processing unit while the DSN processing unit continues to process previously pending data access requests. The method continues with the load balancing module cancelling selection of the DSN processing unit to process the data access request; and receiving a second indication from the DSN processing unit indication that the DSN processing unit is available.

A system, method, and machine-readable storage medium for providing a set of recommended quality of service (QoS) settings are provided. In some embodiments, providing the recommendation includes receiving a set of QoS settings of a volume for a client, a set of measured QoS metrics of the volume for the client, and a measure of load for a slice service corresponding to the volume. Providing the recommendation further includes determining a predicted QoS metric of the volume and a predicted load of the slice service. Providing the recommendation also includes determining, based on the predicted QoS metric, the predicted load, and the set of QoS settings, a set of recommended QoS settings to the client. The set of QoS settings of the volume for the client is then updated with the set of recommended QoS settings.

A method for offloading services of a sewer application in a network system. The method includes receiving, by a first in-network computation offload instance, a first request packet from a client application, wherein the first request packet includes a first application payload for processing by the server application; generating, by the first instance, a modified request packet that includes the first application payload and first offload information that describes the first instance for use by the server application in coordinating offloading processing to one or more in-network computation offload instances; and transmitting, by the first instance, the modified request packet to the next device in the traffic flow between the client application and the server application, wherein the next device is either (1) a second in-network computation offload instance in the traffic flow between the client application and the server application or (2) the sewer application.

A method for offloading services of a sewer application in a network system. The method includes receiving, by a first in-network computation offload instance, a first request packet from a client application, wherein the first request packet includes a first application payload for processing by the server application; generating, by the first instance, a modified request packet that includes the first application payload and first offload information that describes the first instance for use by the server application in coordinating offloading processing to one or more in-network computation offload instances; and transmitting, by the first instance, the modified request packet to the next device in the traffic flow between the client application and the server application, wherein the next device is either (1) a second in-network computation offload instance in the traffic flow between the client application and the server application or (2) the sewer application.

Techniques are disclosed relating to implementing inter-application communication in a peer-to-peer fashion via a lease-based central registry. The central registry is configured to maintain application-layer routing information and to perform load balancing of the inter-application communication. A first application of a plurality of applications may send a lease request identifying the second application to the central registry. The first application receives a lease response that identifies a specific instance of a second application within a plurality of computing nodes and includes a resource allocation defining one or more limits on inter-application communication that the first application is permitted to perform with the specific instance of the second application. Based at least in part on the lease response, the first application performs inter-application communication directly with the specific instance of the second application subject to the one or more limits without requiring routing of the inter-application communication by another entity.

Methods, systems, and apparatus, including an apparatus for generating clusters of building blocks of compute nodes using an optical network. In one aspect, a method includes receiving request data specifying requested compute nodes for a computing workload. The request data specifies a target n-dimensional arrangement of the compute nodes. A selection is made, from a superpod that includes a set of building blocks that each include an m-dimensional arrangement of compute nodes, a subset of the building blocks that, when combined, match the target n-dimensional arrangement specified by the request data. The set of building blocks are connected to an optical network that includes one or more optical circuit switches. A workload cluster of compute nodes that includes the subset of the building blocks is generated. The generating includes configuring, for each dimension of the workload cluster, respective routing data for the one or more optical circuit switches.

A computer-implemented method for deploying an application between an on-premise server and an off-premise server includes identifying a plurality of nodes in a flow of an application deployed on the on-premise server. The computer-implemented method further includes splitting the flow at the plurality of nodes to form a plurality of sub-flows of the application. The computer-implemented method further includes routing a flow execution workload of the application to the plurality of sub-flows of the application.