A relational approach to monitoring complex systems
ACM Transactions on Computer Systems (TOCS)
A bibliography of parallel debuggers, 1990 edition
ACM SIGPLAN Notices
DPM: A Measurement System for Distributed Programs
IEEE Transactions on Computers
| Hi-index | 0.00 |
Conventional programs have a single stream of execution. Distributed programs expand on this notion, having multiple streams of execution that interact with each other. This expansion increases the complexity of a program''s behavior. Existing program monitors do not provide enough information to deal with all of the problems of a distributed computing environment. This paper is concerned with the development of a programming tool for Berkeley UNIX whose goal is to characterize the performance of distributed programs. The Distributed Programs Monitor (DPM) monitors specifically the interactions between the processes of a distributed program and provides routines to analyze the resulting data. DPM is a tool composed of independent subtools that work together to monitor a distributed program. This paper will present an overview of DPM including a bit of its history and a bit of its experimental use, but it deals primarily with the development of a control process for the monitor. The design and the implementation of this process is described. A major issue of the design addresses how to provide distributed process management in a non- distributed processing environment.