Performance Evaluation for Global Computation

Abstract. Global computing applications co-ordinate distributed computations across widely-dispersed hosts. Such systems present formidable design and implementation challenges to software developers and synchronisation, scheduling and performance problems come to the fore. Complex systems such as these can benefit from the application of highlevel performance analysis methods founded on timed process algebras. In this paper we compare the use of two such approaches, the PEPA nets and EOS methods, illustrating our presentation with the example of modelling Web services.

• Type: Conference
• Year: 2003
• Where: GLOBAL
• Authors: Linda Brodo, Pierpaolo Degano, Stephen Gilmore, Jane Hillston, Corrado Priami
• Keywords: Artificial Intelligence, Co-ordinate Distributed Computations, GLOBAL 2003, Performance Analysis Methods, Timed Process Algebras