Research challenges on online service quality prediction for proactive adaptation

Online quality prediction allows service-oriented systems to anticipate the need for adaptation and thus to prevent the actual occurrence of failures or to mitigate upcoming failures. Such proactive adaptation capabilities are increasingly relevant for future service-oriented systems, which need to cope with limited control over third-party services, as well as rapidly changing usage contexts. Initial, promising results have been achieved for what concerns online quality prediction for service-oriented systems. However, there are many challenging issues remaining that call for concrete solutions. In this paper we present a set of research challenges identified by the research community that may be worth investigating in the coming years.