Browse Prior Art Database

Validating Pattern Placement in a Multi-Location Cloud Computing Environment

IP.com Disclosure Number: IPCOM000236367D
Publication Date: 2014-Apr-22
Document File: 2 page(s) / 56K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a method for validating pattern placement in a multi-location cloud computing environment to ensure that all patterns can be successfully deployed.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 01 of 2

Validating Pattern Placement in a Multi-Location Cloud Computing Environment

Application developers can create patterns describing the structure of applications. Pattern engines exist that can use these patterns to deploy instances of the application as one or more virtual machine instances in a virtualized server environment, such as a private, public, or hybrid cloud computing environment. The

placement engine in the cloud infrastructure is responsible for determining the "best" server(s) onto which to place the virtual machine instances. The "best" servers may vary depending on the deployment policies selected by the users. For instance, a high-availability policy may prefer to distribute the workload across multiple regions, sites and servers, whereas an energy conservation policy may prefer to place the

virtual machine instances on the same server instances at the same site.

Locations where a virtual machine instance in a pattern is deployed may become overloaded and lead to contention for physical resources such as from the Central Processing Unit (CPU). A spike in workload might occur for the virtual machine

itself, or other virtual machines might be running hot. When this happens, contention may occur for a given physical resource at the location.

To alleviate the contention, the system may move one or more virtual machine instances from the current location to other location(s). Where there are a small number of virtual servers is relatively easy; it is possible to make an intelligent placement decision using a centralized placement engine that knows the full states of all physical and virtual resources are known in the cloud. However, this does not scale with the addition of physical resources and deployment of more virtual resources.

When deploying a pattern, a user may select the host aggregates to which a virtual machine instance may be deployed. Users may also select to assign an Internet Protocol (IP) address to each of the network interfaces used by each virtual machine instance. However, it may not be possible to successfully deploy a pattern using the selections made by the user. A method is needed to prevent the deployment from

continuing and failing.

...