APE: An Automated Performance Engineering Process for Software as a

Rolia, Jerry ; Krishnamurthy, Diwakar ; Xu, Min ; Graupner, Sven

HP Development Company

关键词: Benchmarking;    Business Applications;    Performance Sizing;    Optimization;    Software as a Service;   

RP-ID  :  HPL-2008-65

学科分类：计算机科学（综合）

美国|英语

来源: HP Labs

【 摘 要 】

The Software as a Service (SaaS) paradigm is changing the way in which businesses procure software solutions. A service provider hosts the software solution amortizing management and infrastructure costs across the businesses it serves. For complex software offerings, each business may use a given software platform in very different ways. This need for customization can pose performance challenges and risks for those hosting the software as a service. This paper presents an Automated Performance Engineering (APE) process for transaction oriented enterprise applications that supports infrastructure selection, sizing, and performance validation for customized service instances in hosted software environments. It enables the rapid deployment of a customized service instance while lowering performance related risks by automating the creation of a customized performance model and customized benchmark model. Case study results demonstrate the effectiveness of the approach for a TPC-W system. 37 PagesExternal Posting Date: June 7, 2008 [Fulltext].Approved for External PublicationInternal Posting Date: June 7, 2008 [Fulltext]