A Catalog of Performance Measures for Self-Adaptive Systems

  • Maike Bezerra da Silva UFC
  • Carla Bezerra UFC
  • Emanuel Coutinho UFC
  • Paulo Henrique Maia UECE

Resumo


[Context] Self-adaptive systems (SAS) can evaluate their own behavior at runtime and change it when necessary to avoid failures. Many of those systems require a high processing power to carry out their dynamic and complex operations successfully. In this realm, performance arises as an important quality characteristic that has been widely used in the literature to drive the necessary adaptation strategies in SAS to achieve or maintain the QoS requirements. However, few studies evaluate the SAS performance efficiently due to the lack of knowledge on all possible performance measures. Still, there is a variety of measures scattered throughout the body of work in this area. [Objective] Our objective is to create a catalog of performance measures for assessing self-adaptive systems. [Method] To do that, we identified and analyzed the performance measures used in 32 primary studies found through a systematic literature mapping. [Results] As a result, 18 measures have been inserted in the catalog, having most of them addressed the sub-characteristics of time behavior and resource utilization. To validate the catalog, we implemented a subset of measures in a healthcare SAS.
Palavras-chave: Self-adaptive systems, Measures, Performance
Publicado
08/11/2021
DA SILVA, Maike Bezerra; BEZERRA, Carla; COUTINHO, Emanuel; MAIA, Paulo Henrique. A Catalog of Performance Measures for Self-Adaptive Systems. In: SIMPÓSIO BRASILEIRO DE QUALIDADE DE SOFTWARE (SBQS), 20. , 2021, Evento Online. Anais [...]. Porto Alegre: Sociedade Brasileira de Computação, 2021 . p. 1-10.