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

Method and Apparatus of Optimal Service Selection for Workflow Execution

IP.com Disclosure Number: IPCOM000016403D
Original Publication Date: 2002-Dec-29
Included in the Prior Art Database: 2003-Jun-21
Document File: 4 page(s) / 83K

Publishing Venue

IBM

Abstract

With the proliferation of the Internet and the wide acceptance of e-commerce, increasing numbers of business processes and services are offered by distributed and heterogeneous service providers. This has created the need to explicitly employ workflow management systems (WFMS) to compose, coordinate and control the flows of services. Since each service offering differs from another along multiple QoS attributed such cost, process time, and reliability, the selection of these services to execute workflow poses a challenge. Currently, most of service selection tools regard tasks in a workflow are independent, without considering constraints (e.g. precedence constraints) among them. In this project, we proposal OptFlow, a framework that optimizes the service selection for whole workflow instance execution. Based on user configurable linear models, it considers constraints among the tasks and selects the optimal combination of service offering to execute the workflow. We adopt linear programming as means to select services. It can guarantee that the selection result is optimal and computation cost is polynomial time.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 53% of the total text.

Page 1 of 4

Method and Apparatus of Optimal Service Selection for Workflow Execution

Introduction

The automation of Web services composition using workflow technologies gaining a considerable momentum as a paradigm for Business-to-Business (B2B) collaboration. Leading organizations are embracing Web services today, realizing the enormous competitive advantages since this model provides faster time to market reducing cycle times and increasing customer service. Widely available and standardized Web services make it possible to realize just-in-time Business-to-Business Integration (B2Bi) which focuses on integrating Web services into a composite web service to create added value. For example, a composite web service can provide a set of high level financial management services which uses payroll, tax preparation, and cash management service as components. Using WFMSs, virtual enterprises might be formed by integrating Web services from various service providers to manufacture and sell the product.

By Web service, also called e-service, (in the remainder, we will use the terms e-service, Web service and service interchangeably), we mean a semantically well defined abstraction that allows user to access functionality offered by Web applications
(i.e. service providers). Each service is provided by multiple providers and differ from another along multiple QoS attributes such as cost, processing time, and reliability. The selection of these services to execute workflow poses a challenge. Currently, most of the tools focus on selecting service for single task, without considering the constraints among the tasks in a workflow instance. Such solutions are easy to implement, but they cannot guarantee that the execution of whole workflow instance is optimal. In fact, an important aspect of executing a workflow instance is the selection of service providers that optimizes the utility measure (such as time to product introduction in market) of the whole workflow instance, subject to considerations such as precedence constraints, processing time, budget, and reliability.

In this article, we outline a suite of decision models and mathematical programming based solution approaches for optimal composition of services to execute workflow instances. We adopt linear programming as means to select services. It can guarantee that the selection result is optimized and computation cost is polynomial time. We have implemented this decision support solution within a Java platform that provides a standard API to work with any workflow related platform.

Fundamental Concepts

In this section, we introduce some basic taxonomy and concepts. Currently, there is no established standard to measure the quality of Web services. We use the Web service quality (WSQ) as a set of quality criteria (see Table 1).

1

Page 2 of 4

Table 1: Web service quality criteria

In OptFlow, state chart diagram is ad...