Scalable virtual network switch architecture
Embodiments of an overlay network with innovative virtual network switches and related network interface controllers are disclosed. According to one embodiment, the overlay network may comprise at least one virtual network switch (VNS), wherein each VNS is installed within a corresponding first processing endpoint which is logically above a virtualization layer of a virtual network. And, the VNS may be configured to intercept data packets transmitted on a network layer logically below the virtualization layer of the virtual network. The VNS may comprise: (a) a manager element configured to connect to at least one virtual network interface controller (VNIC), each VNIC being installed within a corresponding second processing endpoint of the virtual network, the second processing endpoint being separate from the first processing endpoint; (b) a replicator node configured to replicate and route the data packets; and (c) a switch controller configured to control the first replicator node.
1. An overlay network comprising:
at least one virtual network switch (VNS), each VNS being installed within a corresponding first processing endpoint which is logically above a virtualization layer of a virtual network, the VNS being implemented as at least one application running on at least one virtual machine and configured to intercept and switch data packets transmitted on or above the data link layer (L2) of traditional OSI model, thereby performing packet switching among the virtual network of virtual machines and above a physical layer,
wherein the VNS comprises:
(a) a first manager element configured to connect to at least one virtual network interface controller (VNIC), each VNIC being installed within a corresponding second processing endpoint which is also logically above the virtualization layer of the virtual network, the second processing endpoint being separate from the first processing endpoint;
(b) a first replicator node configured to replicate and route the data packets; and
(c) a first switch controller configured to control the first replicator node.
2. The overlay network of claim 1 , wherein each VNS comprises a network virtual appliance configured to execute over one or more virtual machines.
3. The overlay network of claim 2 , wherein the at least one VNS is configured to be scaled by adding or removing virtual machines over which the VNS executes, thereby respectively increasing or decreasing a speed and/or a capacity of the overlay network.
4. The overlay network of claim 1 , wherein the at least one VNS comprises multiple VNSs which are configured to form a mesh.
5. The overlay network of claim 1 , wherein each VNIC connects the corresponding second processing endpoint to the VNS.
6. The overlay network of claim 5 , wherein each VNIC facilitates data packet delivery to and/or retrieval from the corresponding second processing endpoint.
7. The overlay network of claim 1 , wherein the VNS comprises a plurality of replicator nodes including the first replicator node.
8. The overlay network of claim 7 , wherein the plurality of replicator nodes are dynamically assembled into a tree configuration in response to a change in demand for communication speed and/or capacity of the VNS.
9. The overlay network of claim 1 , wherein the first processing endpoint comprises at least one virtual machine, software container, serverless compute function, or physical hardware.
10. The overlay network of claim 1 , wherein each VNS is configured to filter, modify, or capture the data packets or to provide in-switch computing functionality.
11. A method for operating an overlay network, the overlay network comprising at least one virtual network switch (VNS), each VNS being installed within a corresponding first processing endpoint which is logically above a virtualization layer of a virtual network, the method comprising:
connecting, by a first manager element of the VNS, to at least one virtual network interface controller (VNIC), each VNIC being installed within a corresponding second processing endpoint which is also logically above the virtualization layer of the virtual network, the second processing endpoint being separate from the first processing endpoint, wherein the VNS is implemented as at least one application running on at least one virtual machine and configured to intercept and switch data packets transmitted on or above the data link layer (L2) of traditional OSI model, thereby performing packet switching among the virtual network of virtual machines and above a physical layer;
replicating, by a first replicator node of the VNS, data packets received from the at least one VNIC; and
routing, by the first replicator node, the data packets according to control by a first switch controller of the VNS.
12. The method of claim 11 , wherein each VNS comprises a network virtual appliance configured to execute over one or more virtual machines.
13. The method of claim 12 , further comprising:
scaling the at least one VNS by adding or removing one or more virtual machines over which the VNS executes, thereby respectively increasing or decreasing a speed and/or a capacity of the overlay network.
14. The method of claim 11 , wherein the at least one VNS comprises multiple VNSs which are configured to form a mesh.
15. The method of claim 11 , wherein each VNIC connects the corresponding second processing endpoint to the VNS.
16. The method of claim 15 , wherein each VNIC facilitates data packet delivery to and/or retrieval from the corresponding second processing endpoint.
17. The method of claim 11 , wherein the VNS comprises a plurality of replicator nodes including the first replicator node.
18. The method of claim 17 , wherein the plurality of replicator nodes are dynamically assembled into a tree configuration in response to a change in demand for communication speed and/or capacity of the VNS.
19. The method of claim 11 , wherein the first processing endpoint comprises at least one virtual machine, software container, serverless compute function, or physical hardware.
20. The method of claim 11 , wherein each VNS is configured to filter, modify, or capture the data packets or to provide in-switch computing functionality.