Auto-Discovering Configurations for Service Management |
| |
Authors: | Pankaj K Garg Martin Griss Vijay Machiraju |
| |
Affiliation: | (1) HP Labs, 1501 Page Mill Road, Palo Alto, California, 94304 |
| |
Abstract: | Determining service configurations is essential for effective service management. In this paper we describe a model-driven approach for service configuration auto-discovery. We develop metrics for performance and scalability analysis of such auto-discovery mechanisms. Our approach addresses several problems in auto-discovery: specification of what services to discover, how to efficiently distribute service discovery, and how to match instances of services into related groups. We use object-oriented models for discovery specifications, a flexible bus-based architecture for distribution and communication, and a novel multi-phased, instance matching approach. We have applied our approach to typical e-commerce services, Enterprise Resource Planning applications, like SAP, and Microsoft Exchange services running on a mixture of Windows and Unix platforms. The main contribution of this work is the flexibility of our models, architecture and algorithms to address discovery of a multitude of services. |
| |
Keywords: | Service discovery service model performance scalability service components grouping |
本文献已被 SpringerLink 等数据库收录! |
|