作者: Pengcheng Xiong , Zhikui Wang , Simon Malkowski , Qingyang Wang , Deepal Jayasinghe
关键词: Server 、 Workload 、 Service-oriented architecture 、 Service-level agreement 、 Web application 、 Distributed computing 、 Provisioning 、 Resource allocation 、 Control reconfiguration 、 Computer science 、 Cloud computing 、 Resource management
摘要: Resource provisioning for N-tier web applications in Clouds is non-trivial due to at least two reasons. First, there an inherent optimization conflict between cost of resources and Service Level Agreement (SLA) compliance. Second, the resource demands multiple tiers can be different from each other, varying along with time. Resources have allocated (virtual) containers minimize total amount while meeting end-to-end performance requirements application. In this paper we address these challenges through combination controllers on both application container levels. On level, a decision maker (i.e., adaptive feedback controller) determines budget that are required meet SLA as workload varies. second controller partitions among components optimize round trip time). We evaluated our method three models -- open, closed, semi-open were implemented RUBiS benchmark. Our evaluation indicates major advantages comparison previous approaches. fewer provisioned achieve same performance. approach robust enough various types workloads time-varying demand without reconfiguration.