As a guest user you are not logged in or recognized by your IP address. You have
access to the Front Matter, Abstracts, Author Index, Subject Index and the full
text of Open Access publications.
In SaaS cloud, there are two interrelated steps to schedule user requests. The first is to choose the proper VM (service) to handle the request, and the second is to map each VM (service) onto a proper physical machine (server). In this paper, we first put forward a heuristic algorithm to enhance the resource utilizations in a heterogenous data center and then propose a cascaded fuzzy dynamic programming (CFDP) method to do joint optimization of tasks and VM migrations to pursue long-term and maximum rewards with multi-objectives, which refer to performance benefits, energy consumptions, cooling costs for hot-spots, penalties for unfinished requests before deadline and VM migration costs. CFDP can learn from characteristics of incoming requests, does not rely on priori prediction of request arrivals, and considers both current and future impacts in the process of decision-making. Besides, value function approximation method is used to cope with state explosions and guarantee high scheduling speed at the same time. Finally, a group of experiments with both synthetic and realistic workloads are conducted, and QoS comparisons verify the effectiveness and robustness of our approach. In theory, we have extended the ability to comprehensively make two or more correlated decisions together in long-term and multi-objective optimizations with extremely large shared state.
This website uses cookies
We use cookies to provide you with the best possible experience. They also allow us to analyze user behavior in order to constantly improve the website for you. Info about the privacy policy of IOS Press.
This website uses cookies
We use cookies to provide you with the best possible experience. They also allow us to analyze user behavior in order to constantly improve the website for you. Info about the privacy policy of IOS Press.