Browse Prior Art Database

Dynamic Deployment of Large Scale Virtual Machines in Cloud Environment

IP.com Disclosure Number: IPCOM000245664D
Publication Date: 2016-Mar-28
Document File: 5 page(s) / 66K

Publishing Venue

The IP.com Prior Art Database

Abstract

There is a need to detect and manage the virtual network to dynamic deploy the new virtual machines in the cloud computing environment, the disclosure introduces a new way for Monitor to collect the network information and provide the solution about the optimal virtual machines deployment in the virtual network, make sure the new deployed virtual machine could communicate with the network and it will also not cause the the congestion in the existing network.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 46% of the total text.

Page 01 of 5

Dynamic Deployment of Large Scale Virtual Machines in Cloud Environment

With the development of cloud computing, network planning and deployment are getting more and more attention. In cloud infrastructure, there are plenty of resources, such as processors, memory and input/output devices, some of them could be coupled as a subset, and acted as an independent server. These resources may be dedicated to a particular subset or shared by multiple subsets. Each subset runs as if it is executing on its own dedicated machine even though the physical resources are virtualized and may be assigned from a shared pool of resources. For this reason, a subset here is referred to as a virtual machine (VM). Some of the VMs are linked to the same virtual switch and communicate with each other. These VMs could also communicate with the physical network which is connected through the virtual switch. There maybe are hundreds of virtual machines running on one hypervisor.These virtual machines are linked with plenty of virtual switches. So there will be a need to detect and manage the virtual network to dynamic deploy the new virtual machines. Currently, if the user couldn't get the whole network status, there is a possibility that the new virtual machine is attached with a very busy virtual switch which will cause the network congestion finally. The disclosure will focus on the optimal virtual machines deployment in the virtual network, make sure the new deployed virtual machine could communicate with the network and it will also not cause the the congestion in the existing network.

This disclosure introduces a new way to deploy new virtual machines based on the current network status: Dynamic Virtual Machine Deployment.

It takes the bottom+to+top solution to tell the centralized control system (which is called the monitor in this disclosure ) to get network traffic status, which means the virtual node (virtual switch or virtual machine) takes the initiative to send the report to the monitor. It requires to implement a traffic statistics program in each virtual node, to count the number of transmitted packets, received packets, dropped packets. Then the virtual node sends these information to the monitor periodically. In this way, monitor could get the traffic information in the network and figure out a best virtual switch to connect when deploying a new virtual machine. Besides, there are some popular technique solutions to keep the new deployed virtual machine communicate with the designated network, such as vxlan, gre, etc. no more tautology here.

This disclosure has the following advantages:

It provides a more efficient manner to deploy virtual machines, which would enable better utilization of the virtual switches.

The current network congestion solutions focus on the migration of network, it may cause a virtual machine to be migrated immediately once deployed. This disclosure considers the network status when deploying a new virtual machine,...