We present a model-based performance testing approach using the MBPeT tool. We use of probabilistic timed automata to model the user profiles and to generate synthetic workload. The MBPeT generates the load in a distributed fashion and applies it in real-time to the system under test, while measuring several key performance indicators, such as response time, throughput, error rate, etc. At the end of the test session, a detailed test report is provided. MBPeT has a distributed architecture and supports load generation distributed over multiple machines. New generation nodes are allocated dynamically during load generation. In this book chapter, we will present the MBPeT tool, its architecture, and demonstrate its applicability with a set of experiments on a case study. We also show that using abstract models for describing the user profiles allows us quickly experiment different load mixes and detect worst case scenarios.
|Title of host publication
|Developing Cloud Software : algorithms, applications, and tools
|I Porres, T Mikkonen, A Ashraf
|TUCS general publications
|Published - 2013
|MoE publication type
|A3 Part of a book or another research book