Browse Prior Art Database

Deriving SLA for Composite Web Services with Multiple Implemnetation Options

IP.com Disclosure Number: IPCOM000175762D
Original Publication Date: 2008-Oct-23
Included in the Prior Art Database: 2008-Oct-23

Publishing Venue

IBM

Abstract

Web services can be aggregated to create composite workflows (i.e. composite web services) that provide streamlined functionality for human users or other systems. For both web service consumers and providers, a Service Level Agreement (SLA) is an important means to ensure the quality-of-service of the invoked web service. In many scenarios, an SLA is made available in the form of response time guarantees along with the expected cost to consumers and penalty to providers. However, it is not straight forward to derive SLA for a composite web service that is built on top of other web services, each with their own SLA. Furthermore, a composite web service can and typically has multiple implementations for the purpose of fault tolerance and high flexibility. In this paper we focus on the problem of deriving the SLA of a composite web service workflow with multiple implementation options. We show that a genetic search algorithm is appropriate to map composite web service requests to the underlying implementations, service type instances, and servers for optimal load distribution. We also develop a methodology to systematically generating a feasible SLA for a composite service to its full capability. Through experimentation we show that our algorithm and methodology scale well up to a large scale system configuration comprising up to 1000 work°ow requests to a complex composite web services with multiple implementations.

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 8% of the total text.

Page 1 of 13

Deriving SLA for Composite Web Services with Multiple Implemnetation Options

Background: Web services can be aggregated to create composite workflows (i.e. composite web services) that provide streamlined functionality for human users or other systems. For both web service consumers and providers, a Service Level Agreement (SLA) is an important means to ensure the quality-of-service of the invoked web service. In many scenarios, an SLA is made available in the form of response time guarantees along with the expected cost to consumers and penalty to providers. However, it is not straight forward to derive SLA for a composite web service that is built on top of other web services, each with their own SLA. Furthermore, a composite web service can and typically has multiple implementations for the purpose of fault tolerance and high flexibility.

Summary: In this disclosure we focus on the problem of deriving the SLA of a composite web service workflow with multiple implementation options. We show that a genetic search algorithm is appropriate to map composite web service requests to the underlying implementations, service type instances, and servers for optimal load distribution. We also develop a methodology to systematically generating a feasible SLA for a composite service to its full capability. Through experimentation we show that our algorithm and methodology scale well up to a large scale system configuration comprising up to 1000 workflow requests to a complex composite web services with multiple implementations.

1

Page 2 of 13

Deriving SLA for Composite Web Services with Multiple Implementation Options

Thomas Phan ¤

     Yahoo!, Inc.
2821 Mission College Blvd. Santa Clara, CA 95054, USA

thomas.phan@acm.org

Wen-Syan Li

IBM Almaden Research Center 650 Harry Road
San Jose, CA, 95120, USA

wsl@us.ibm.com

ABSTRACT

Web services can be aggregated to create composite work- fLows (i.e. composite web services) that provide streamlined functionality for human users or other systems. For both web service consumers and providers, a Service Level Agree- ment (SLA) is an important means to ensure the quality-of- service of the invoked web service. In many scenarios, an SLA is made available in the form of response time guaran- tees along with the expected cost to consumers and penalty to providers. However, it is not straight forward to derive SLA for a composite web service that is built on top of other web services, each with their own SLA. Furthermore, a com- posite web service can and typically has multiple implemen- tations for the purpose of fault tolerance and high fLexibility. In this paper we focus on the problem of deriving the SLA of a composite web service workfLow with multiple imple- mentation options. We show that a genetic search algo- rithm is appropriate to map composite web service requests to the underlying...