Distributed computing system with hypervisor mobility service

A method implemented at a distributed computing system, the method comprising:

• at a first mobility client implemented by a first hypervisor executing on a first computing device,
• detecting an attachment of a virtual machine to the first hypervisor;
• responsive to detecting the attachment of the virtual machine to the first hypervisor, incrementing a first mobility sequence number;
• and
• sending to a mobility service a notification of the attachment of the virtual machine to the first hypervisor;
• at a second mobility client implemented by a second hypervisor executing on a second computing device,
• detecting an attachment of the virtual machine to the second hypervisor;
• sending to the first computing device a message indicating the attachment of the virtual machine to the second hypervisor;
• responsive to detecting the attachment of the virtual machine to the second hypervisor, incrementing a second mobility sequence number;
• and
• sending to the mobility service a notification of the attachment of the virtual machine to the second hypervisor;
• determining, by a tracking service, whether the virtual machine is attached to the first hypervisor or the second hypervisor based on determining which of the first mobility sequence number and the second mobility sequence number is larger;
• and
• at the first mobility client, when the virtual machine is determined to be attached to the second hypervisor,
• based at least on the message, storing network routing information indicating a network location of the virtual machine as attached to the second hypervisor;
• receiving network traffic for the virtual machine that was routed to the first hypervisor at the first computing device;
• and
• forwarding the network traffic for the virtual machine from the first computing device to the second hypervisor implemented on the second computing device based on the stored network routing information.

A distributed computing system, comprising:

• a server executing a mobility service;
• a first computing device executing a first hypervisor implementing a first mobility client configured to:
• detect an attachment of a virtual machine to the first hypervisor;
• responsive to detecting the attachment of the virtual machine to the first hypervisor, increment a first mobility sequence number;
• and
• send to the mobility service a notification of the attachment of the virtual machine to the first hypervisor;
• a second computing device executing a second hypervisor implementing a second mobility client configured to:
• detect an attachment of the virtual machine to the second hypervisor;
• send to the first computing device a message indicating the attachment of the virtual machine to the second hypervisor;
• responsive to detecting the attachment of the virtual machine to the second hypervisor, increment a second mobility sequence number;
• and
• send to the mobility service a notification of the attachment of the virtual machine to the second hypervisor,
• a software-defined network configured to:
• receive network traffic for the virtual machine routed to the first hypervisor at the first computing device, when the virtual machine is attached to the second hypervisor;
• and
• based on software-defined network routing information indicating the attachment of the virtual machine to the first hypervisor, route the network traffic to the first computing device,
• wherein the first mobility client is further configured to:
• based at least on the message, store network routing information indicating a network location of the virtual machine as attached to the second hypervisor, the stored network routing information including a virtual machine identifier identifying the virtual machine and an endpoint internet protocol (IP) address of the second computing device;
• receive network traffic for the virtual machine that was routed to the first hypervisor at the first computing device;
• and
• forward the network traffic for the virtual machine from the first computing device to the second hypervisor implemented on the second computing device based on the stored network routing information, and
• wherein a tracking service determines whether the virtual machine is attached to the first hypervisor or the second hypervisor based on determining which of the first mobility sequence number and the second mobility sequence number is larger.