Juniper Networks, Inc.

Nutanix, Inc.

Hexin Wang

Igor Grobman

Gregory A Smith

Xiang Zhu

Xiangwen Xu

In one embodiment, a system for migrating virtual machines in a virtualization environment includes a plurality of host machines implementing the virtualization environment and a migration controller. Each of the host machines includes a hypervisor, one or more user virtual machines (UVMs) and a virtual machine controller. The system further implements a virtual disk comprising a plurality of storage devices, the virtual disk being accessible by the virtual machine controllers, which conduct I/O transactions with the virtual disk. The migration controller determines a segment size and, for each host machine, determines a number of required segments for the UVMs running on the host machine. The controller computes a number of reserved segments based on a total number of host machines and a largest one of the numbers of required segments. The reserved segments are then assigned among the plurality of host machines.

Handling migration in a virtualization environment

A system and method for reserving resources for a virtual machine to use during failure of an underlying node are disclosed. The system and method include a scheduling module that is configured to reserve resources for the virtual machine on at least one candidate node. To reserve the resources, the scheduling module is configured to identify a list of candidate nodes based upon a qualifier function, rank the candidate nodes based upon a priority function, and determine the candidate nodes that satisfy a high availability threshold. The scheduling module is also configured to select a highest-ranked candidate node that satisfies the high availability threshold and reserve the resources on the highest-ranked candidate node.

Maintaining high availability during N-node failover

In one embodiment, a method includes receive a translation vector, selecting a translation entry from a plurality of translation entries, and determining whether the translation entry is associated with a first identifier class or a second identifier class. The translation vector includes a first identifier, a second identifier, and a virtual memory identifier. The first identifier is associated with a first identifier class, and the second identifier is associated with a second identifier class. The translation vector is received from a translation module including a memory configured to store the plurality of translation entries. Each translation entry from the plurality of translation entries including a virtual memory identifier. The translation entry is selected from the plurality of translation entries of the translation module based on the virtual memory identifier of the translation vector. The determining whether the translation entry is associated with the first identifier class or the second identifier class is based on a shared indicator associated with the translation entry.

Methods and apparatus for improved access to shared memory

In some embodiments, a non-transitory processor-readable medium includes code to cause a processor to receive, at a network management module, a request for data plane information associated with a set of access switches of a distributed switch. The non-transitory processor-readable medium includes code to cause the processor to send, in response to the request, an instruction to each access switch from the set of access switches such that a proxy module at each access switch accesses data plane information at at least one line card at that access switch. The non-transitory processor-readable medium includes code to cause the processor to receive, from each access switch from the set of access switches, the data plane information associated with that access switch, and then send a signal to output, on a single interface, the data plane information associated with each access switch from the set of access switches.

Growing community of inventors

Los Altos, CA, United States of America

Hexin Wang