Applying SPE techniques for modeling a grid-enabled JAVA platform

Advances in Internet and the availability of powerful computers and high-speed networks have propitiated the rise of Grids. The scheduling of applications is complex in Grids due to factors like the heterogeneity of resources and changes in their availability. Performance models provide a way of performing repeatable and controllable experiments for evaluating scheduling algorithms under different scenarios. This article describes the development of a performance model for a JAVA based distributed platform using a SPE methodology. Our case study is SUMA, a distributed execution platform implemented on top of Grid services. It was necessary the use of Software Performance Engineering techniques for understanding and modeling the system. We applied a software performance methodology where Layered Queuing Network (LQN) models are derived from Use Case Maps (UCM). At the end we obtained a performance model of SUMA and a better-documented system. Categories and Subject Descriptors C.4 [Co...

• Type: Conference
• Year: 2005
• Where: WOSP
• Authors: Mariela Curiel, M. Angélica Pérez, Ricardo González
• Keywords: Performance Model, Software Engineering, Software Performance, Software Performance Engineering, WOSP 2005