Managing distributed systems performance and capacity: a three-level approach

摘要

This paper presents an approach to distributed systems performance and capacity management. It begins with a definition of the goals of distributed systemns performance and capacity management. it then explores three levels of the effort, which are: Discrete Element Management, process Management and Application Management. It discusses the metrics that are available in each level, and the comparative benefits of each approach. Finally, it identifies some functional capabilities that distributed systems management tools should have in order to meet the goals initially presented. Examples are drawn from a variety of distributed platforms, including Windows NI, UNIX and MVS.