AutoParam : Automated Control of Application-Level Performance in Virtualized Server Environments

Configuring virtual machine parameters in a virtualized server environment is a challenging task for IT operators. For instance, while current management products can dynamically size virtual machines in order to maintain resource utilization targets, default target values are rarely ideal for individual workloads and operators find it difficult to decide appropriate values for the targets. This paper presents AutoParam, a tool that provides application-level performance guarantees by automatically determining system-level parameters such as the utilization targets and the sizes of the virtual servers hosting individual tiers of multi-tier applications. AutoParam is based on synthesis of a feed-forward transaction-mix-based queueing model and feedback control loops. We describe the integration of AutoParam with the Xen virtualization system, and present empirical results showing that AutoParam effectively adjusts sizes of Xen virtual machine containers to maintain mean transaction response time at a desired level in spite of variable workloads. We compared AutoParam to four other designs, and show that it provides more reasonable tradeoffs between application performance and resource efficiency as well as more robust dynamic properties.

[1]  Sang Hyuk Son,et al.  A feedback control approach for guaranteeing relative delays in Web servers , 2001, Proceedings Seventh IEEE Real-Time Technology and Applications Symposium.

[2]  Chenyang Lu,et al.  An adaptive control framework for QoS guarantees and its application to differentiated caching , 2002, IEEE 2002 Tenth IEEE International Workshop on Quality of Service (Cat. No.02EX564).

[3]  K. Shin,et al.  Performance Guarantees for Web Server End-Systems: A Control-Theoretical Approach , 2002, IEEE Trans. Parallel Distributed Syst..

[4]  C. Amza,et al.  Specification and implementation of dynamic Web site benchmarks , 2002, 2002 IEEE International Workshop on Workload Characterization.

[5]  Lui Sha,et al.  Queueing model based network server performance control , 2002, 23rd IEEE Real-Time Systems Symposium, 2002. RTSS 2002..

[6]  S. Parekh,et al.  MIMO control of an Apache web server: modeling and controller design , 2002, Proceedings of the 2002 American Control Conference (IEEE Cat. No.CH37301).

[7]  HarrisTim,et al.  Xen and the art of virtualization , 2003 .

[8]  Lui Sha,et al.  Feedback control with queueing-theoretic prediction for relative delay guarantees in web servers , 2003, The 9th IEEE Real-Time and Embedded Technology and Applications Symposium, 2003. Proceedings..

[9]  T. Abdelzaher,et al.  Improved prediction for Web server delay control , 2004, Proceedings. 16th Euromicro Conference on Real-Time Systems, 2004. ECRTS 2004..

[10]  Yixin Diao,et al.  Feedback Control of Computing Systems , 2004 .

[11]  Xiaoyun Zhu,et al.  Utilization and SLO-Based Control for Dynamic Sizing of Resource Partitions , 2005, DSOM.

[12]  Terence Kelly,et al.  Detecting Performance Anomalies in Global Applications , 2005, WORLDS.

[13]  Lui Sha,et al.  Modeling 3-tiered Web applications , 2005, 13th IEEE International Symposium on Modeling, Analysis, and Simulation of Computer and Telecommunication Systems.

[14]  Asser N. Tantawi,et al.  An analytical model for multi-tier internet services and its applications , 2005, SIGMETRICS '05.

[15]  T. Kelly,et al.  Predicting Performance in Distributed Enterprise Applications , 2006 .

[16]  Christopher Stewart,et al.  Exploiting nonstationarity for performance prediction , 2007, EuroSys '07.