Browse Prior Art Database

Method to increase application availability through parallel processing of resources

IP.com Disclosure Number: IPCOM000200109D
Publication Date: 2010-Sep-28
Document File: 5 page(s) / 39K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a method to increase application availability through parallel processing of resources

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 38% of the total text.

Page 01 of 5

Method to increase application availability through parallel processing of resources

A software program that keeps applications highly available in a clustered environment is responsible for bringing required resources online to support the successful operation of one or more applications, with the end goal of keeping the supported applications highly available. Keeping applications highly available means that they are started and fully operational and accessible by the end user as much as possible, with less than some small percentage of overall down time.

When two or more applications are being brought online by such a high availability application, there may be dependencies between them such that one application (the parent) must be started before the second one (the child) can be started. In this case in current products that are responsible for bringing such applications to the operating state, the following sequence is followed (serially):
1. All resources required by the parent application are first brought online,
2. The parent application is started,
3. Then all resources required by the child application are made available,
4. And finally the child application is started.

Application resources can include (but are not limited to) disks, volume groups, file systems, IP addresses, tape and other storage devices, NFS mounts and exports, and custom user defined resources.

The time to bring on a parent-child application is the sum of the time to complete all of these steps, and they must be completed in the order shown for both applications to be started successfully. This time can be significant in some cases, decreasing application availability (when an application is being brought online, it is unavailable until all steps are complete and both the parent and child application are operational). The most time is generally spent bringing resources online, particularly when there are many resources required to support an application.

This method proposes a way to significantly reduce the time required to bring ordering (parent-child) dependent applications online, in order to improve application availability. When considering a group of dependent applications in a high availability cluster, this method can significantly improve fallover time (the time to move applications from one cluster node to another when necessary).

This method proposes a way to improve application availability by decreasing the time required to bring ordering dependent applications online (such as parent-child dependent applications). This is accomplished by processing resources for multiple applications in parallel. This is possible with applications that have ordering dependencies because the dependency requirement is typically for one application to start before another starts, and not that a parent application's resources are available before the child's resources.

Consider the simplest case with two applications. One is the parent application and th...