Managing external communications for provided computer networks

• US 8,645,508 B1
• Filed: 03/03/2010
• Issued: 02/04/2014
• Est. Priority Date: 03/03/2010
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Dickinson, Andrew B., Cohn, Daniel T.
• Primary Examiner(s): Etienne, Ario
• Assistant Examiner(s): Williams, Clayton R
• Application Number:
  US12/717,001
• Time in Patent Office: 1,434 Days
• Field of Search: None
• US Class Current: 709/220
• CPC Class Codes: H04L 45/64 using an overlay routing layer
• CPC Class Codes: H04L 45/7453 using hashing
• CPC Class Codes: H04L 67/025 for remote control or remote monitoring of applications
• CPC Class Codes: H04L 67/1023 based on a hash applied to IP addresses or costs
• CPC Class Codes: H04L 67/1097 for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]
• CPC Class Codes: H04L 67/59 Providing operational support to end devices by off-loading in the network or by emulation, e.g. when they are unavailable
• CPC Class Codes: H04L 9/0643 Hash functions, e.g. MD5, SHA, HMAC or f9 MAC

Abstract

Techniques are described for providing managed computer networks. In some situations, the techniques include managing communications for computing nodes of a managed computer network by using a pool of multiple alternative intermediate destinations to forward at least some communications to other final destinations. For example, a manager module associated with a source computing node may select a particular one of multiple alternative intermediate destinations to use for one or more particular communications from the source computing node to an indicated final destination, such as based on network addresses associated with the source computing node and/or the final destination. The multiple alternative intermediate destinations may have various forms, including multiple alternative edge devices at a first location at which various computing nodes are co-located that operate to manage communications between those co-located computing nodes at the first location and other external computer systems at one or more other locations.

Claims

1. A computer-implemented method comprising:

providing, by one or more programmed computing systems, a first virtual computer network by overlaying the first virtual computer network on a distinct substrate network;

receiving an indication of a first communication that is directed to a first network address from a first computing node of the first virtual computer network at a first geographical location, the first network address being associated with a first external destination that is not at the first geographical location;

before forwarding the first communication over the substrate network from the first computing node, selecting, by the one or more programmed computing systems, an edge device from a group of multiple alternative edge devices connected to the substrate network at the first geographical location to use for the first communication, the selecting being performed in a configured manner that is based at least in part on the first network address; and

initiating forwarding, by the one or more programmed computing systems, of the first communication over the substrate network from the first computing node to the selected edge device, to enable further forwarding of the first communication by the selected edge device to the first external destination.

2. The method of claim 1 wherein each edge device of the group has a distinct associated substrate network address, and wherein the forwarding of the first communication over the substrate network address to the selected edge device is performed using the substrate network address associated with the selected edge device.

3. The method of claim 2 wherein the edge devices of the group are associated with a block of multiple substrate network addresses that includes the substrate network address associated with the selected edge device, and wherein the selecting of the edge device includes performing hashing using the first network address that results in selecting the substrate network address associated with the selected edge device from the block.

4. The method of claim 3 wherein the first computing node is associated with a second virtual network address for the first virtual computer network, and wherein the performing of the hashing is further based in part on the second virtual network address.

5. The method of claim 4 wherein the first external destination is a second computing node of the first virtual computer network that is located at a second geographical location distinct from the first geographical location, and wherein the first network address is a virtual network address of the first virtual computer network that is associated with the second computing node.

6. The method of claim 3 wherein the performing of the hashing is further based in part on one or more attributes associated with the first communication that are distinct from the first network address and from a second virtual network address associated with the first computing node.

7. The method of claim 4 wherein the first computing node is one of multiple virtual machines hosted on a physical computing system, wherein each of the multiple virtual machines has a distinct associated network address, wherein the distinct network address associated with the first computing node virtual machine is the second virtual network address for the first virtual computer network, and wherein the selecting of the edge device is performed by a virtual machine hypervisor module that executes on the physical computing system to manage communications by the multiple virtual machines.

8. The method of claim 4 wherein the first computing node is one of multiple virtual machines hosted on a physical computing system, and wherein the second virtual network address for the first virtual computer network is associated with the physical computing system.

9. The method of claim 4 wherein the first network address associated with the first external destination is a substrate network address.

10. The method of claim 1 further comprising, before forwarding a second communication from a second computing node of the first virtual computer network to a second external destination, selecting an edge device from the group in a random manner for use in forwarding the second communication to the second external destination.

11. The method of claim 1 further comprising tracking a current load on at least one of the edge devices of the group, and wherein the selecting of the edge device is further performed based at least in part on the tracked current load.

12. The method of claim 1 further comprising, after the forwarding of the first communication over the substrate network to the selected edge device:

receiving one or more indications of one or more second communications sent from the first computing node to the first external destination; and

for at least one of the one or more second communications, selecting an edge device from the group that is distinct from the edge device selected for use with the first communication, and forwarding the second communication over the substrate network to the selected edge device distinct from the edge device selected for use with the first communication, to enable spreading communications from the first computing node to the first external destination across two or more of the multiple alternative edge devices.

13. The method of claim 1 further comprising, after the forwarding of the first communication over the substrate network to the selected edge device:

receiving an indication of a second communication sent from the first computing node to a second external destination that is not at the first geographical location, the second external destination being indicated by a second network address distinct from the first network address;

selecting a second of the edge devices of the group to use based at least in part on the second network address, the selected second edge device being distinct from the selected edge device for use with the first communication; and

forwarding the second communication over the substrate network to the selected second edge device for further forwarding by the selected second edge device to the second external destination.

14. The method of claim 13 further comprising, after the forwarding of the first communication over the substrate network to the selected edge device:

receiving an indication of a third communication sent from a second computing node at the first geographical location to the first network address, the second computing node being associated with a third virtual network address for the first virtual computer network;

selecting the second edge device to use for the third communication based at least in part on the first network address and on the third virtual network address; and

forwarding the third communication over the substrate network to the selected second edge device for further forwarding by the selected second edge device to the first external destination.

15. The method of claim 1 further comprising, after the forwarding of the first communication over the substrate network to the selected edge device:

receiving one or more indications of one or more second communications that are each sent from the first computing node to the first external destination using the first network address;

selecting the edge device to use for each of the one or more second communications; and

forwarding each of the one or more second communications over the substrate network to the selected edge device for further forwarding by the selected edge device to the first external destination.

16. The method of claim 15 wherein the selecting of the edge device to use for the first communication includes tracking an association of the selected edge device with a combination of the first computing node and the first external destination, and wherein the selecting of the edge device to use for each of the one or more second communications is based on the tracked association of the selected edge device with the combination of the first computing node and the first external destination.

17. The method of claim 15 wherein the selecting of the edge device to use for the first communication includes selecting the selected edge device after the receiving of the indication of the first communication based on first information associated with the first communication, and wherein, for each of the one or more second communications, the selecting of the edge device to use for the second communication includes dynamically selecting the selected edge device based on the first information being associated with the second communication.

18. The method of claim 1 wherein the first virtual computer network includes multiple computing nodes at the first geographical location that each has an associated substrate network address corresponding to a location of the computing node in the substrate network, and wherein the method further comprises forwarding one or more additional communications over the substrate network address to a destination computing node of the multiple computing nodes by using the substrate network address associated with the destination computing node.

19. The method of claim 18 wherein the multiple computing nodes each further has an associated virtual network address for the first virtual computer network that is independent from the associated substrate network address for that computing node.

20. The method of claim 1 wherein the one or more programmed computing systems are part of a configurable network service that provides the first virtual computer network to a first client and that provides multiple other virtual computer networks to multiple other clients, and wherein the providing of the first virtual computer network is performed in accordance with received configuration information for the first virtual computer network by emulating functionality of one or more indicated networking devices of the first virtual computer network without physically providing the indicated networking devices.

21. A non-transitory computer-readable storage medium having stored contents that configure a computing system to:

obtain information regarding a first virtual computer network having multiple computing nodes, the first virtual computer network being overlaid on a distinct substrate network that interconnects the multiple computing nodes;

manage a communication that is from a first computing node of the multiple computing nodes, wherein the first computing node is provided by the configured computing system, and wherein the communication has a destination network address associated with an external final destination that is not part of the first virtual computer network, the managing of the communication including;

before the communication leaves the configured computing system, selecting one intermediate destination from multiple alternative intermediate destinations in the substrate network to use in forwarding the communication to the external final destination, the selecting of the one intermediate destination being performed in a configured manner that is based at least in part on the destination network address for the communication; and

determining a substrate network address that is associated with the selected one intermediate destination; and

initiate the forwarding of the communication over the substrate network from the configured computing system to the selected one intermediate destination using the determined substrate network address, to enable further forwarding of the communication by the selected one intermediate destination to the external final destination.

22. The non-transitory computer-readable storage medium of claim 21 wherein the multiple computing nodes of the first virtual computer network are co-located at a first location, wherein the multiple alternative intermediate destinations in the substrate network are multiple alternative edge devices that manage communications between computing nodes at the first location and external computer systems at other locations, and wherein each of the multiple alternative edge devices has a distinct substrate network address that is used when forwarding a communication over the substrate network to that alternative edge device.

23. The non-transitory computer-readable storage medium of claim 21 wherein the stored contents are executable software instructions that further configure the computing system to, for a second communication sent to a destination virtual network address associated with a second of the multiple computing nodes, initiate forwarding of the second communication over the substrate network to a location of the second computing node in the substrate network by using a second substrate network address associated with the second computing node.

24. A computing system, comprising:

one or more hardware processors; and

a manager module that is associated with one or more computing nodes that are at least some of multiple computing nodes of a first computer network overlaid on one or more substrate second networks, wherein the manager module is configured to, when executed by at least one of the one or more hardware processors, manage a communication sent by a first computing node of the one or more computing nodes by;

before the communication is sent over the one or more substrate second networks, selecting an intermediate destination from multiple alternative intermediate destinations in the one or more substrate second networks to use in forwarding the communication to an external final destination, the selecting of the intermediate destination being performed in a configured manner that is based at least in part on a destination network address specified for the communication;

determining a substrate network address for the one or more substrate second networks that is associated with the selected intermediate destination; and

sending the communication from the first computing node to the one or more substrate second networks for forwarding to the selected intermediate destination by using the determined substrate network address.

25. The computing system of claim 24 wherein the multiple computing nodes of the first computer network are located at one or more first locations, wherein the external final destination is at a location remote from the one or more first locations, and wherein the multiple alternative intermediate destinations are multiple alternative edge devices that manage communications between computing nodes at the one or more first locations and external computer systems at other locations.

26. The computing system of claim 24 wherein the manager module includes software instructions for execution by the at least one processors, wherein the first computer network is a virtual computer network, and wherein each of the multiple computing nodes has an associated virtual network address for the first computer network and has a distinct associated substrate network address for the one or more substrate second networks.

27. The computing system of claim 24 wherein the manager module is further configured to manage a second communication sent by the first computing node by:

determining that a second destination network address for the second communication is associated with a second external final destination that is not one of the multiple computing nodes, the second destination network address for the second communication being distinct from the destination network address for the communication;

selecting a second intermediate destination from the multiple alternative intermediate destinations to use in forwarding the second communication to the second external final destination, the selecting of the second intermediate destination being based at least in part on the second network address, the first and second intermediate destinations being distinct; and

sending the second communication from the first computing node to the one or more substrate second networks for forwarding to the selected second intermediate destination using a second substrate network address associated with the selected second intermediate destination.