Method and apparatus to replicate stateful virtual machines between clouds

摘要

A method including replicating a first virtual machine (VM) in a first cloud and putting the replicated VM in a second cloud. Activating the first VM and pausing the replicated VM. First processing, at the first VM, traffic from VMs in the first cloud, wherein the first processing occurs when the first VM is activated and the replicated VM is paused. Buffering, at a hypervisor of the replicated VM, traffic from VMs in the second cloud, wherein the buffering occurs when the first VM is activated and the replicated VM is paused. Activating the replicated VM in response to state information of the first VM and pausing the first VM. Second processing, at the replicated VM, the buffered traffic according to the state information of the first VM, wherein the second processing occurs when the replicated VM is activated and the first VM is paused.

主权项

1. A method, comprising:
replicating a first virtual machine (VM) found in a first cloud computing environment and putting the replicated VM in a second cloud computing environment; wherein the first cloud computing environment comprises: a first hypervisor coupled to the first VM; a first flip engine included with the first hypervisor; and a first buffer included with the first hypervisor for buffering traffic of VMs within the first cloud computing environment destined for the first VM; wherein the second cloud computing environment comprises: a second hypervisor coupled to the replicated VM; a second flip engine included with the second hypervisor; and a second buffer included with the second hypervisor for buffering traffic of VMs within the second cloud computing environment destined for the replicated VM; activating the first VM and pausing the replicated VM; first processing, at the first VM, traffic from the VMs in the first cloud computing environment, wherein the first processing occurs when the first VM is activated and the replicated VM is paused; buffering, at the second buffer of the second hypervisor corresponding to the replicated VM, traffic destined for the replicated VM from the VMs in the second cloud computing environment, wherein the buffering occurs when the first VM is activated and the replicated VM is paused; determining, with the first flip engine, state information for sending a state delta signal to the replicated VM for activation, wherein the state information represents a number of packets received at the first VM or a number of packets stored in the second buffer of the replicated VM; automatically sending the state delta signal from the first flip engine to the second flip engine, wherein the state delta signal includes virtual memory and central processing unit states that have been updated since the first VM became active; in response to the state delta signal, memory pages modified by the first VM overwrite state memory pages in the replicated VM before the replicated VM becomes active so that the replicated VM processes the buffered traffic in the second buffer at the last state of the first VM; activating the replicated VM in response to overwriting state memory pages and pausing the first VM; and second processing, at the replicated VM, the buffered traffic in the second buffer according to the state information of the first VM, wherein the second processing occurs when the replicated VM is activated and the first VM is paused.