Part VIII Concluding Remarks 1
Learning Objectives Summarize what we have learned in this tutorial Present and comment a bibliography for capacity planning and performance analysis of Web servers. 2
Web and Intranet Performance: a quantitative analysis This tutorial gave the participants the foundations required to carry out capacity planning and performance analysis studies of Web-based systems. The main steps are based on two models - a workload model and a performance model that were discussed in detail in this tutorial. Ó 1998 Menascé & Almeida. All Rights Reserved. 3
Summary Web Performance Problems Internet and Web growth Examples of Web performance problem situations Anatomy of an HTTP request Performance aspects of TCP and HTTP Ó 1998 Menascé & Almeida. All Rights Reserved. 4
Summary Queuing Theory and Operational Analysis service time and service demand waiting time and queuing time Basic Performance Results and Examples utilization law forced flow law service demand law Little’s Law 5
Summary Web Performance Issues Combination of HTTP and TCP/IP Simple examples using operational analysis Bottlenecks Perception of performance and metrics Quality of Service Web caching proxy Ó 1998 Menascé & Almeida. All Rights Reserved. 6
Summary Capacity Planning Methodology Concept of adequate capacity Service Level Agreement (SLA) Framework of a methodology for capacity planning: workload characterization workload forecasting performance modeling and prediction model validation and cost model 7
Summary Workload Characterization Methodology Choice of an analysis standpoint Identification of the basic component Choice of the characterizing parameters Data collection Partitioning the workload Calculating the class parameters Averaging Clustering techniques and algorithms 8
Summary New Phenomena in the Internet and WWW Burstiness Heavy-tailed distributions Data Collection Measurement techniques Web server access log 9
Summary System Models for the Web System-level models view a server as a black box. Only its arrival process and throughput functions are relevant. State Transition Diagrams (STDs) can be used to find the probability that k requests are in the server. Use the flow in = flow out principle. Little’s Law can be used to compute the response time from the average number of requests and from the throughput. 10
Summary Component-level models Parameters for component-level models include the service demands on system resources, i. e. total time spent by a request receiving service from a resource. Waiting times, response times, throughputs can be computed using open models (e. g. , web servers) closed models (e. g. , intranets) 11
Bibliography • Capacity Planning and for Web Performance: metrics, models and methods, Daniel Menascé and Virgilio Almeida, Prentice Hall, Upper Saddle River, 1998. • Capacity Planning and Performance Modeling: from mainframes to client/server systems, Daniel Menascé, Virgilio Almeida, and Larry Dowdy, Prentice Hall, Upper Saddle River, 1994. • Practical Planning of Network Growth, John Blommers, Hewlett Packard, Prentice Hall, Upper Saddle River, 1996. • The Art of Computer Systems Performance Analysis: techniques for experimental design, measurement, simulation and modeling, Raj Jain, John Wiley & Sons, 1991 12
Bibliography • The Benchmark Handbook for Database and Transaction Processing System, 2 nd Edition, Jim Gray, Morgan Kaufmann, 1993. • Queueing Systems, Leonard Kleinrock, John Wiley & Sons, 1975. • Web Server Technology, N. Yeager and R. Mc. Grath, Morgan Kaufmann, San Francisco. 1996. • “Measuring the Capacity of a Web server”, G. Banga and P. Druschel, USENIX Symposium on Internet Technology and Systems, Dec. 1997. 13
Bibliography • “Self-Similarity in World Wide Web Traffic: evidence and possible causes”, M. Crovella and A. Bestravos, Proceedings of the 1996 SIGMETRICS Conf. Measurement Comput. Syst. , ACM, Philadelphia, May 1996. • “Measuring the Behavior of a World Wide Web server”, J. Almeida, V. Almeida, D. Yates, Proc. 7 th Conf. High Perform. Networking (HPN 97), IFIP, New York, April 1997. • “Web Workload Characterization”, M. Arlitt and G. Williamson, Proceedings of the 1996 SIGMETRICS Conf. Measurement Comput. Syst. , ACM, Philadelphia, May 1996. • “A Framework for Software performance Engineering of client/server systems”, D. Menascé, Proc. 1997 Computer Measurement Group Conference, Orlando Dec. 1997. 14
Скачать презентацию Part VIII Concluding Remarks 1 Learning Objectives
d44ba43e552579c3c9cadea0f4a7c3d5.ppt