Systems and methods to create highly scalable network services
Various embodiments of the invention provide for an unparalleled scalable and reliable load distribution design for use in a large class of network applications, including cloud computing in modern data center networks and similar services, which can benefit from the invention by taking advantage of commonly available network components and systems. In particular, various embodiments of the invention extend a logical load distributor model to include distributed clusters of loosely coupled components. Certain embodiments accomplish this by combining existing silicon-based network components and functionalities with the scalability and reliability provided by virtualized processes in order to improve critical load distribution and load balancing performance without requiring costly upgrades to infrastructure and without negative structural impact on the overall network. Some embodiments take advantage of hierarchical designs to further increase scalability.
1. An information handling system for communicating with one or more virtual network engines, the information handling system comprising:
one or more network interfaces communicatively connected to a network fabric that supports a plurality of virtual connections to a plurality of instances of a network engine operating on one or more hosts, the one or more hosts communicatively connected to the network fabric;
one or more processors that are communicatively coupled to the one or more network interfaces;
a non-transitory memory comprising one or more sequences of instructions that, when executed by the one or more processors, cause steps to be performed comprising forming a virtual connection to each of the plurality of instances of the network engine, the virtual connection facilitating data traffic between the information handling system and the instance of the network engine; and
a load balancing module, communicatively coupled to the one or more processors, that performs load balancing of the data traffic across at least some of the plurality of virtual connections.
2. The information handling system according to claim 1 wherein the non-transitory memory further comprises one or more sequences of instructions which, when executed by the one or more processors, causes steps to be performed comprising:
responsive to a new instance of the network engine being created on the one or more hosts, forming a new virtual connection to the new instance of the network engine; and
responsive to the instance of the network engine ceasing to operate, removing the instance of the network engine.
3. The information handling system according to claim 1 , wherein each instance of the network engine in the plurality of instances of network engines emulate a network layer pairing.
4. The information handling system according to claim 3 wherein the network layer pairing is at Layer 3 and the plurality of instances of network engines form separate logical units regardless of whether one or more of the instances of the network engine operate on a same host, and the plurality of virtual connections form an Equal Cost Multi-Path.
5. The information handling system according to claim 3 wherein the network layer pairing is at Layer 2 and the plurality of virtual connections form a link aggregation group.
6. The information handling system according to claim 5 wherein two or more of the plurality of instances of network engines form a single logical entity.
7. The information handling system according to claim 1 , wherein two or more of the plurality of instances of the network engine operate on one of the one or more hosts.
8. An information handling system for communicating with virtual network engines, the information handling system comprising:
one or more hosts connected to a network fabric that is communicatively connected to one or more network interfaces; and
a plurality of instances of a network engine operating on the one or more hosts, the network fabric supporting a plurality of virtual connections between a network device and each of the plurality of instances of the network engine to receive data traffic from the network device, the virtual connections being formed in response to one or more sequences of instructions held in a non-transitory memory being executed by one or more processors that are communicatively coupled to the one or more network interfaces and to a load balancing module that performs load balancing of the data traffic across at least some of the plurality of virtual connections.
9. The information handling system according to claim 8 , wherein each instance of the network engine in the plurality of instances of network engines emulate a network layer pairing.
10. The information handling system according to claim 9 wherein the network layer pairing is at Layer 2 and the plurality of virtual connections form a link aggregation group.
11. The information handling system according to claim 10 wherein the network layer pairing is at Layer 3 and the plurality of instances of network engines form separate logical units regardless of whether one or more of the instances of the network engine operate on a same host, and the plurality of virtual connections form an Equal Cost Multi-Path.
12. The information handling system according to claim 8 , wherein the plurality of instances of the network engine is configured to communicate with each other and one of a network node and a proxy for a network comprising the network node via a distributed message bus using a messaging protocol to distribute control messages.