Predictive analytics for Smarttarget: Proactive performance anomaly detection techniques for cloud-based Web services

Smarttarget 的预测分析：基于云的 Web 服务的主动性能异常检测技术

• 批准号: 514610-2017
• 负责人: Krishnamurthy, Diwakar
• 金额: $ 1.82万
• 依托单位: University of Calgary
• 依托单位国家: 加拿大
• 项目类别: Engage Grants Program
• 财政年份: 2017
• 资助国家: 加拿大
• 起止时间: 2017-01-01 至 2018-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=640681
• 关键词: Predictive; analytics; Smarttarget; Proactive; performance

The private sector partner for this research project is Ottawa based Predikat Inc. The company provides a suiteof application performance management (APM) tools called Smarttarget for their customers. Predikat'scustomers typically host Web services on cloud platforms such as Amazon EC2 and Microsoft Azure. Webservice performance is an important consideration for these customers. In particular, customers would like toensure fast response times for their Web services to support good Quality of Experience (QoE) to end users ofthese services. Unfortunately, several factors make it challenging to achieve good performance. In particular,the workload experienced by a Predikat Web service can often fluctuate significantly. Furthermore, cloudplatforms are known to introduce performance degradations since in such platforms multiple applications cancompete for shared cloud resources. Consequently, a Predikat customer requires APM tools that canautomatically and proactively predict the impending onset of performance problems. Such a capability wouldallow the customer to quickly put in place strategies to mitigate the problem thereby ensuring that end userQoE is not compromised. The purpose of this Engage collaboration is to identify the challenges in buildingsuch tools and to investigate effective solutions for these challenges.

这个研究项目的私营部门合作伙伴是位于渥太华的Predikat公司。该公司为其客户提供了一套名为smartttarget的应用程序性能管理（APM）工具。Predikat的客户通常将Web服务托管在亚马逊EC2和微软Azure等云平台上。对于这些客户来说，web服务性能是一个重要的考虑因素。特别是，客户希望确保其Web服务的快速响应时间，以支持这些服务的最终用户获得良好的体验质量（QoE）。不幸的是，有几个因素使其难以实现良好的性能。特别是，Predikat Web服务所经历的工作负载通常波动很大。此外，众所周知，云平台会带来性能下降，因为在这样的平台上，多个应用程序可以竞争共享的云资源。因此，Predikat客户需要能够自动主动预测即将出现的性能问题的APM工具。这样的功能将允许客户快速制定策略来缓解问题，从而确保最终用户的qoe不会受到损害。此次Engage合作的目的是确定构建此类工具的挑战，并研究应对这些挑战的有效解决方案。