Configuring Workload Manager Control Parameters for Resource Pools

Rolia, Jerome ; Cherkasova, Ludmila ; McCarthy, Clifford

HP Development Company

关键词: workload managers;    automation;    enterprise applications;    shared resource pools;    QoS metrics;    QoS;    measurements;    capacity management;    performance models;   

RP-ID  :  HPL-2006-57

学科分类：计算机科学（综合）

美国|英语

来源: HP Labs

【 摘 要 】

Resource pools are computing environments that offer virtualized access to shared resources. When used effectively they can align the use of capacity with business needs (flexibility), lower infrastructure costs (via resource sharing), and lower operating costs (via automation). Using resources effectively can rely on a combination of workload placement and workload management technologies. Workload placement decides which workloads will share resources. Workload management governs short term access to resource capacity. It provides performance isolation within resource pools to ensure resource sharing even under high loads. A workload manager can have a direct impact both on an application's overall resource access quality of service and on the number of workloads that can be assigned to a pool. In this paper we take a detailed look at an application workload's demands. We explore tradeoffs in resource access quality of service received by the application and the minimum allocation of resources for the workload. We show that by careful selection of workload scheduling parameters along with a proposed fast allocation policy we can sometimes more than triple the number of workloads that can be assigned to a pool without sacrificing application workload quality of service or the efficiency of the resource pool. Notes: Copyright IEEE 2006. Published in the Proceedings of the 2006 IEEE/IFIP Network Operations & Management Symposium (NOMS 2006), 3-7 April 2006, Vancouver, Canada 11 Pages