This paper discusses the design andperformance of a hierarchical proxy-cache designed to make Internet information systems scale better. The design was motivated by our earlier trace-driven simulation study of Internet traffic. We believe that the conventional wisdom, that the benefits of hierarchical file caching do not merit the costs, warrants reconsideration in the Internet environment. The cache implementation supports a highly concurrent stream of requests. We present performance measurements that show that the cache outperforms other popular Internet cache implementations by an order of magnitude under concurrent load. These measurements indicate that hierarchy does not measurably increase access latency. Our software can also be configured as a Web-server accelerator; we present data that our httpd-accelerator is ten times faster than Netscape's Netsite and NCSA 1.4 servers. Finally, we relate our experience fitting the cache into the increasingly complex and operational world of Internet information systems, including issues related to security, transparency to cache-unaware clients, and the role of file systems in support of ubiquitous wide-area information systems.

Abstract – The Internet is rapidly growing in number of users, traffic levels, and topological complexity. At the same time it is increasingly driven by economic competition. These developments render the characterization of network usage and workloads more difficult, and yet more critical. Few recent studies have been published reporting Internet backbone traffic usage and characteristics. At MCI, we have implemented a high-performance, low-cost monitoring system that can capture traffic and perform analyses. We have deployed this monitoring tool on OC-3 trunks within internetMCI’s backbone and also within the NSF-sponsored vBNS. This paper presents observations on the patterns and characteristics of wide-area Internet traffic, as recorded by MCI’s OC-3 traffic monitors. We report on measurements from two OC-3 trunks in MCI’s commercial Internet backbone over two time ranges (24-hour and 7-day) in the presence of up to 240,000 flows. We reveal the characteristics of the traffic in terms of packet sizes, flow duration, volume, and percentage composition by protocol and application, as well as patterns seen over the two time scales. 1

The phenomenal growth in popularity of the World Wide Web (WWW, or the Web) has made WWW traffic the largest contributor to packet and byte traffic on the NSFNET backbone. This growth has triggered recent research aimed at reducing the volume of network traffic produced by Web clients and servers, by using caching, and reducing the latency for WWW users, by using improved protocols for Web interaction. Fundamental to the goal of improving WWW performance is an understanding of WWW workloads. This paper presents a workload characterization study for Internet Web servers. Six different data sets are used in this study: three from academic environments, two from scientific research organizations, and one from a commercial Internet provider. These data sets represent three different orders of magnitude in server activity, and two different orders of magnitude in time duration, ranging from one week of activity to one year of activity. Throughout the study, emphasis is placed on finding wor...

Abstract not found

The widespread use of the World Wide Web and related applications places interesting performance demands on network servers. The ability to measure theeffect of these demands is important for tuning and optimizing the various software components that make up a Web server. To measure these effects, it is necessary to generate realistic HTTP client requests. Unfortunately, accurate generation of such traffic in a testbed of limited scope is not trivial. In particular, the commonly used approach is unable to generate client request-rates that exceed the capacity of the server being tested even for short periods of time. This paper examines pitfalls that one encounters when measuring Web server capacity using a synthetic workload. We propose and evaluate a new method for Web traffic generation that can generate bursty traffic, with peak loads that exceed the capacity of the server. Finally, we use the proposed method to measure the performance of a Web server.

Is it possible to reduce the expected response time ofevery request at a web server, simply by changing the order in which we schedule the requests? That is the question we ask in this paper. This paper proposes a method for improving the performance of web servers servicing static HTTP requests. The idea is to give preference to those requests which are short, or have small remaining processing requirements, in accordance with the SRPT (Shortest Remaining Processing Time) scheduling policy. The implementation is at the kernel level and in-volves controlling the order in which socket buffers are drained into the network.Experiments are executed both in a LAN and a WAN environment. We use the Linux operating system and the Apache and Flash web servers. Results indicate that SRPT-based scheduling of connections yields significant reductions in delay at the web server. These result in a substantial reduction inmean response time, mean slowdown, and variance in response time for both the LAN and WAN environments. Significantly, and counter to intuition, the large requests are only negligibly penalized or not at all penalized as a result of SRPT-based scheduling.

Abstract—With the recent explosion in usage of the World Wide Web, the problem of caching Web objects has gained considerable importance. Caching on the Web differs from traditional caching in several ways. The nonhomogeneity of the object sizes is probably the most important such difference. In this paper, we give an overview of caching policies designed specifically for Web objects and provide a new algorithm of our own. This new algorithm can be regarded as a generalization of the standard LRU algorithm. We examine the performance of this and other Web caching algorithms via event- and trace-driven simulation.

To date there have been a number of efforts that attempt to characterize various aspects of the World Wide Web. This paper presents a summary of these efforts, highlighting regularities and invariants that have been discovered. Keywords: Statistics, Metrics, Analysis, and Modeling

A distributed multi-server Web site can provide the scalability necessary to keep up with growing client demand at popular sites. Load balancing of these distributed Web-server systems, consisting of multiple Web servers for document retrieval and a Domain name server (DNS) for address resolution, opens interesting new problems. In this paper, we investigate the effects of using a more active DNS which, as an atypical centralized scheduler, applies some scheduling strategy in routing the requests to the most suitable Web server. Unlike traditional parallel/distributed systems in which a centralized scheduler has full control of the system, the DNS controls only a very small fraction of the requests reaching the multi-server Web site. This peculiarity, especially in the presence of highly skewed load, makes it very difficult to achieve acceptable load balancing and avoid overloading some Web server. This paper adapts traditional scheduling algorithms to the DNS, proposes new policies, a...

Most well-managed web servers perform well most of the time. Occasionally, however, every popular web server experiences transient overload. An overloaded web server typically displays signs of its affliction within a few seconds. Work enters the web server at a greater rate than the web server can complete it, causing the number of connections at the server to build up. This implies large delays for clients accessing the server. This paper provides a systematic performance study of exactly what happens when a web server is run under transient overload, both from the perspective of the server and from the perspective of the client. Second, this paper proposes and evaluates a particular kernel-level solution for improving the performance of web servers under overload. The solution is based on SRPT connection scheduling. We show that SRPT-based scheduling improves overload performance across a variety of client and server-oriented metrics.