Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Intelligently perform workload splits and pattern / product migrations for adaptive growth

IP.com Disclosure Number: IPCOM000236038D
Publication Date: 2014-Apr-02
Document File: 4 page(s) / 46K

Publishing Venue

The IP.com Prior Art Database

Abstract

This article describes the system that provides a decision making engine to manage the performance and growth of an application by migrating a workload from one type of pattern to another type of pattern. An enterprise can choose the pattern of their choice for an application, and when an application workload reaches a certain threshold, the engine makes the decision about the pattern components that should be replaced. This system determines intelligently what services to split and how to split out service onto their own VMs. System stores a set of policies to upgrade based on best practice models. These policies helps in decision process.

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

Page 01 of 4

Intelligently perform workload splits and pattern / product migrations for adaptive growth

In any organization, most of the applications start as a pilot or prototype and go through production deployment lifecycle. As the adoption of an application workload grows, the underlying architecture also needs to evolve.

In cloud deployment model and PaaS (Platform as a Service) deployment model, the decision to change the underlying middleware infrastructure can be pre-planned, automated and optimized. Workload configurations can be migrated from one topology to another, based on the several factors. Performance, business criticality and SLAs to deliver the services are some of the factors and these factors change over the life-span of an application.

Described in this article is the system that provides a decision making engine to manage the performance and growth of an application by migrating a workload from one type of pattern to another type of pattern. An enterprise can choose the pattern of their choice for an application, and when an application workload reaches a certain threshold, the engine makes the decision about the pattern components that should be replaced. This system leverages the known concept of patterns to start with consolidated services on single VM and determines intelligently what services to split and how to split out service onto their own VMs .The existing PaaS system will be leveraged to manage and upgrade the application middlewares. System also stores a set of policies to upgrade based on best practice models. These policies helps in decision process.

The system defines a method to describe the growth policy for an application. It consists of base decision checkpoints; application components performance monitoring and rules engine to make decisions.

High Level Examples:

Example One:

In the following scenario, take the deployment which starts out as a consolidated deployment all on one server. This application is initially used only for US companies which has less strict data protection law. Over time however additional users are added include some from over seas. This however triggers a compliance issue which dictates that international data must be isolated and secured on dedicated resources. This triggers threshold #1, splitting the database onto its own dedicated node.

As the application continues to grow, availability becomes a concern. It is at this point, threshold #2 where the underlying technical products must be upgraded to provide this capability. In this example, WAS is updated to WAS NG which provides a cluster web application topology and DB2* is upgraded to include HADR capabilities. Additionally the components are all split out onto their own dedicated resources to reduce coupling. In this way this application has been able to adapt to many different changing demands automatically.

1


Page 02 of 4

Example Two:

Seen inthis second example, the workload starts out with standard pattern that ant...