부산대학교 도서관

As cloud computing systems are reaching the stagewhere the human effort required to maintain them at anoperational level is unsupportable, one of the major challengesfaced by cloud computing providers is to develop appropriatemechanisms for run-time monitoring and adaptation, to preventcloud platforms from quickly dissolving into a non-reliableenvironment. In this context, the application of intelligentapproaches to Autonomic Clouds may offer promisingopportunities. In this paper we present a novel approach toprovide cloud platforms with autonomic capabilities based onutilising techniques from the domains of the Semantic Web andStream Reasoning. The main idea of this approach is to encodevalues, monitored within cloud platforms, with Semantic Weblanguages, which will allow us to integrate these semanticallyenrichedobservation streams with static ontological knowledgeand apply intelligent reasoning. Based on such run-timereasoning capabilities, we are able to perform analysis andfailure diagnosis, and suggest further adaptation actions. As aninitial proof of concept, we sketch out a conceptual architecturefor a self-adaptation framework and introduce a prototypesolution implementing this architecture.