Conference Paper

Performance Modeling of an Apache Web Server with Bursty Arrival Traffic.

Conference: Proceedings of the International Conference on Internet Computing, IC '03, Las Vegas, Nevada, USA, June 23-26, 2003, Volume 2
Source: DBLP


Performance modeling is an important topic in capacity planning and overload control for web servers. We present a queueing model of an Apache web server that uses bursty arrival traffic. The arrivals of HTTP requests is assumed to be a Markov Modulated Poisson Process and the service discipline of the server is processor sharing. The total number of requests that can be processed at one time is limited to K. We obtain web server performance metrics such as average response time, throughput and blocking probability by simulations. Compared to other models, our model is conceptually simple. The model has been validated through measurements and simulations in our lab. The per- formance metrics predicted by the model fit well to the experimental outcome. Keywords--Internet, World Wide Web, web server, perfor- mance model, MMPP.

Download full-text


Available from: Maria Kihl,
  • Source
    • "Its prediction capability relies on the estimation of the request rates to the corresponding objects. Approaches exist in the literature to provision resources accounting for workload variations [1], [18], however models are not learned from real traffic which restricts predictive capabilities. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Resource allocation in the cloud is usually driven by performance predictions, such as estimates of the future incoming load to the servers or of the quality-of-service(QoS) offered by applications to end users. In this context, characterizing web workload fluctuations in an accurate way is fundamental to understand how to provision cloud resources under time-varying traffic intensities. In this paper, we investigate the Markovian Arrival Processes (MAP) and the related MAP/MAP/1 queueing model as a tool for performance prediction of servers deployed in the cloud. MAPs are a special class of Markov models used as a compact description of the time-varying characteristics of workloads. In addition, MAPs can fit heavy-tail distributions, that are common in HTTP traffic, and can be easily integrated within analytical queueing models to efficiently predict system performance without simulating. By comparison with traced riven simulation, we observe that existing techniques for MAP parameterization from HTTP log files often lead to inaccurate performance predictions. We then define a maximum likelihood method for fitting MAP parameters based on data commonly available in Apache log files, and a new technique to cope with batch arrivals, which are notoriously difficult to model accurately. Numerical experiments demonstrate the accuracy of our approach for performance prediction of web systems.
    IEEE International Conference on Cloud Computing, CLOUD 2011, Washington, DC, USA, 4-9 July, 2011; 01/2011
  • Source
    • "AutoTune agent framework based on intelligent agents and control theory for automatic performance tuning of a Web server [7], and queuing models to study the behavior of the Apache Web server as in [1]. "
    [Show abstract] [Hide abstract]
    ABSTRACT: There is an increasing need for new Web server architectures that are application-centric, simple, small, and pervasive in nature. In this pap er, we present a novel architecture for a bare PC Web server that meets most of these requirements. A bare PC Web server runs on any Intel 386 (or above) based architecture with no operating system in the machine. The application object or program contains all the necessary code t o run in the bare machine including its boot and load programs. This approach to building Web servers has many notable features including optimized task scheduling, zero copy memory buffering, minimal interrupt intervention, concurrency control avoidance, minimal resource dependence, and potential scalability via a cluster of bare PC Web servers. The bare PC Web server executable is small, self-contained, and fits on a single floppy disk. The server does not u se a local hard disk and the resource files are located on the network. We compare performance measurements for the bare PC Web server with three other optimized commercial Web servers: Microsoft IIS, Apache, and Tux. It is found that the bare PC Web server achieves an average fir st response time that is 3.9 - 5.4 times better, and a connection time that is 1.4 - 2.8 times bette r than the other servers. We also discuss alternative designs for bare PC Web servers and com pare their performance.
  • Source
    • "For smaller implementations, the experimental results obtained in the modeling of Apache Web Server [17] should prove helpful. The work done in [17] describes a model of the Apache web server, which consists of a processor sharing node with a queue attached to it. The total number of jobs in the system is limited. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Wireless communication technologies like GPRS, UMTS and WLAN, combined with the availability of high-end, affordable mobile devices enable the development of advanced and innovative mobile services. Devices such as mobile phones and personal digital assistants let the users access a wide range of new offerings whenever and wherever they happen to be. A strategic approach for the quality assurance of these mobile data services should take into account a number of characteristics unique to the mobile paradigm such as the increased complexity of emerging handheld devices, the greater sensitivity to security and load related problems in wireless infrastructure and increased complexities of scale. This paper identifies the major factors influencing the development and testing strategies for these applications and works out effective quality assurance principles to ensure productive and scalable mobile data services
    Communication System Software and Middleware, 2006. Comsware 2006. First International Conference on; 01/2006
Show more