Remote profiling of resource constraints of web servers using mini-flash crowds

作者: Balachander Krishnamurthy , Anees Shaikh , Aditya Akella , Vyas Sekar , Pratap Ramamurthy

DOI:

关键词:

摘要: Unexpected surges in Web request traffic can exercise server-side resources (e.g., access bandwidth, processing, storage etc.) undesirable ways. Administrators today do not have requisite tools to understand the impact of such "flash crowds" on their servers. Most servers either rely over-provisioning and admission control, or use potentially expensive solutions like CDNs, ensure high availability face flash crowds. A more fine-grained understanding performance individual server under emulated but realistic controlled crowd-like conditions aid administrators make efficient resource management decisions. In this paper, we present miniflash crowds (MFC) - a light-weight profiling service that reveals bottlenecks infrastructure. MFC uses set probes where an increasing number distributed clients synchronized requests specific portions remote server. We carried out lab-based tests experiments collaboration with operators production show our approach faithfully track loads different provide useful insights constraints components also results from measurement study provisioning several hundred popular servers, few startup companies, about phishing

参考文章(0)