http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Algorithms for Network-Aware Application Component Placement for Cloud Resource Allocation
Maryam Barshan,Hendrik Moens,Steven Latré,Bruno Volckaert,Filip De Turck 한국통신학회 2017 Journal of communications and networks Vol.19 No.5
Due to the soaring popularity of cloud-based services overthe last years, the size and the complexity of cloud environmentshas been growing quickly. In the context of cloud systems, mappinga number of application components to a set of physical serversand assigning cloud resources to those components is challenging. Traditional resource allocation systems which rely on a centralizedmanagement paradigm suffer from scalability issues, makingthem inappropriate for large-scale cloud environments. Therefore,there is a need for providing new management solutions that scalewell to large size cloud systems. In this article, we present optimaland heuristic solutions for network-aware placement of multicomponentapplications with differing component characteristics. The optimal integer linear programming (ILP)-based solution minimizesthe application rejection rate and the cost of mapping whilerespecting application component requirements and physical networklimitations. As the execution time of the optimal model scalesexponentially, we also offer scalable heuristic solutions for centralizedand hierarchical application placement, which are thoroughlyexplained and evaluated and compared to the optimal solution. Our evaluations show that while the proposed centralized heuristicis near-optimal, the hierarchical approach is much faster and offershigher scalability compared to a centralized approach, e.g., mapping2.7 million application components onto 512k servers. Moreover,the percentage of servers used and fully placed applicationsremain close to that of the centralized and optimal solutions.