Performance Analysis of Client-Server Mass Storage Systems

We consider the performance analysis of a system in which mass storage is provided by a file server node on a network of powerful processors. The server controls a hierarchy of storage devices, and provides all the necessary file management functions. The clients have their own local storage devices whose management is not affected by any of the functional software at the server node. An expert system manages the movement of files between different storage levels. We develop and solve a closed queueing network model of the shared service mass storage system, which is used to analyze the performance of the system for different file movement criteria, and to study the tradeoffs, between recalling a file and accessing it by a distributed file system. Several network topologies and their effect on performance are considered.