The analysis of performance of software architectures has primary importance for many applications. Typical questions that are posed to a performance analyst are:
- Given a certain workload, what will be the response time expected by a user?
- What is the maximum amount of simultaneous users that my system can handle while maintaining a certain quality of service?
- How can I identify the bottleneck of my system?
Performance assessment can be done in three ways:
- By using stochastic simulation
- By resorting to analytical models (such as queueing models)
- By running a benchmarking experiment set
The performance evaluation group within the ACADIA lab is active in the research of new methods for quantitative evaluation of computer systems and applies the techniques at the state of the art for the analysis of wireless sensor networks and distributed software architectures.
ACADIA research tracks related to performance evaluation span three major areas: