USENIX Association Proceedings of USITS ’ 03 : 4 th USENIX Symposium on Internet Technologies and Systems

As Internet services become more popular and pervasive, a critical problem that arises is managing the performance of services under extreme overload. This paper presents a set of techniques for managing overload in complex, dynamic Internet services. These techniques are based on an adaptive admission control mechanism that attempts to bound the 90th-percentile response time of requests flowing through the service. This is accomplished by internally monitoring the performance of the service, which is decomposed into a set of event-driven stages connected with request queues. By controlling the rate at which each stage admits requests, the service can perform focused overload management, for example, by filtering only those requests that lead to resource bottlenecks. We present two extensions of this basic controller that provide class-based service differentiation as well as application-specific service degradation. We evaluate these mechanisms using a complex Web-based e-mail service that is subjected to a realistic user load, as well as a simpler Web server benchmark.

[1]  Joseph L. Hellerstein,et al.  Using Control Theory to Achieve Service Level Objectives In Performance Management , 2001, 2001 IEEE/IFIP International Symposium on Integrated Network Management Proceedings. Integrated Network Management VII. Integrated Management Strategies for the New Millennium (Cat. No.01EX470).

[2]  E. N. Elnozahy,et al.  Measuring Client-Perceived Response Time on the WWW , 2001, USITS.

[3]  William LeFebvre,et al.  CNN.com: Facing a World Crisis , 2001, LiSA.

[4]  Peter Druschel,et al.  A Scalable and Explicit Event Delivery Mechanism for UNIX , 1999, USENIX Annual Technical Conference, General Track.

[5]  David E. Culler,et al.  Overload management as a fundamental service design primitive , 2002, EW 10.

[6]  Prasant Mohapatra,et al.  An Admission Control Scheme for Predictable Server Response Time for Web Accesses , 2001, WWW '01.

[7]  Willy Zwaenepoel,et al.  Flash: An efficient and portable Web server , 1999, USENIX Annual Technical Conference, General Track.

[8]  Ludmila Cherkasova,et al.  Session Based Admission Control: A Mechanism for Improving the Performance of an Overloaded Web Server , 1998 .

[9]  Sugih Jamin,et al.  A measurement-based admission-controlled Web server , 2000, Proceedings IEEE INFOCOM 2000. Conference on Computer Communications. Nineteenth Annual Joint Conference of the IEEE Computer and Communications Societies (Cat. No.00CH37064).

[10]  Sharad Singhal,et al.  Web2K: Bringing QoS to Web Servers , 2000 .

[11]  David E. Culler,et al.  Scalable, distributed data structures for internet service construction , 2000, OSDI.

[12]  Chenyang Lu,et al.  Modeling and performance control of Internet servers , 2000, Proceedings of the 39th IEEE Conference on Decision and Control (Cat. No.00CH37187).

[13]  David E. Culler,et al.  An architecture for highly concurrent, well-conditioned internet services , 2002 .

[14]  Katsuhiko Ogata,et al.  Modern Control Engineering , 1970 .

[15]  Sang Hyuk Son,et al.  Design and evaluation of a feedback control EDF scheduling algorithm , 1999, Proceedings 20th IEEE Real-Time Systems Symposium (Cat. No.99CB37054).

[16]  H. Jamjoom,et al.  QGuard: Protecting Internet Servers from Overload , 2000 .

[17]  Dakshi Agrawal,et al.  Inferring client response time at the web server , 2002, SIGMETRICS '02.

[18]  David E. Culler,et al.  SEDA: an architecture for well-conditioned, scalable internet services , 2001, SOSP.

[19]  Yixin Diao,et al.  Using MIMO feedback control to enforce policies for interrelated metrics with application to the Apache Web server , 2002, NOMS 2002. IEEE/IFIP Network Operations and Management Symposium. ' Management Solutions for the New Communications World'(Cat. No.02CH37327).

[20]  Peter Druschel,et al.  Lazy receiver processing (LRP): a network subsystem architecture for server systems , 1996, OSDI '96.

[21]  Peter B. Danzig,et al.  A Hierarchical Internet Object Cache , 1996, USENIX ATC.

[22]  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.

[23]  Prasant Mohapatra,et al.  Session-based overload control in QoS-aware Web servers , 2002, Proceedings.Twenty-First Annual Joint Conference of the IEEE Computer and Communications Societies.

[24]  Duane Wessels.  Squid internet object cache , 1996 .

[25]  Larry L. Peterson,et al.  Defensive programming , 2002, OSDI.

[26]  Krishna Kant,et al.  Overload Control Mechanisms for Web Servers , 2001 .

[27]  Mark Handley,et al.  Congestion control for high bandwidth-delay product networks , 2002, SIGCOMM.

[28]  Edward W. Knightly,et al.  Multi-class latency-bounded Web services , 2000, 2000 Eighth International Workshop on Quality of Service. IWQoS 2000 (Cat. No.00EX400).

[29]  Srinivasan Keshav.  A control-theoretic approach to flow control , 1991, SIGCOMM 1991.

[30]  Eric A. Brewer,et al.  Cluster-based scalable network services , 1997, SOSP.

[31]  D.E. Tolmie.  Gigabit networking , 1992, IEEE LTS.

[32]  David E. Culler,et al.  Ninja: A Framework for Network Services , 2002, USENIX Annual Technical Conference, General Track.

[33]  Douglas M. Freimuth,et al.  Kernel Mechanisms for Service Differentiation in Overloaded Web Servers , 2001, USENIX Annual Technical Conference, General Track.

[34]  Amin Vahdat,et al.  Differentiated multimedia Web services using quality aware transcoding , 2000, Proceedings IEEE INFOCOM 2000. Conference on Computer Communications. Nineteenth Annual Joint Conference of the IEEE Computer and Communications Societies (Cat. No.00CH37064).

[35]  QUTdN QeO,et al.  Random early detection gateways for congestion avoidance , 1993, TNET.

[36]  Eric A. Brewer,et al.  Harvest, yield, and scalable tolerant systems , 1999, Proceedings of the Seventh Workshop on Hot Topics in Operating Systems.