The Internet is undergoing substantial changes from a communication and browsing infrastructure to a medium for conducting business and marketing a myriad of services. The World Wide Web provides a uniform and widely-accepted application interface used by these services to reach multitudes of clients. These changes place the web server at the center of a gradually emerging eservice infrastructure with increasing requirements for service quality and reliability guarantees in an unpredictable and highly-dynamic environment.

A significant fraction of software failures in large-scale Internet systems are cured by rebooting, even when the exact failure causes are unknown. However, rebooting can be expensive, causing nontrivial service disruption or downtime even when clusters and failover are employed. In this work we separate process recovery from data recovery to enable microrebooting -- a fine-grain technique for surgically recovering faulty application components, without disturbing the rest of the application.

This paper presents a method for admission control and request scheduling for multiply-tiered e-commerce Web sites, achieving both stable behavior during overload and improved response times. Our method externally observes execution costs of requests online, distinguishing different request types, and performs overload protection and preferential scheduling using relatively simple measurements and a straightforward control mechanism. Unlike previous proposals, which require extensive changes to the server or operating system, our method requires no modifications to the host O.S., Web server, application server or database. Since our method is external, it can be implemented in a proxy. We present such an implementation, called Gatekeeper, using it with standard software components on the Linux operating system. We evaluate the proxy using the industry standard TPC-W workload generator in a typical three-tiered e-commerce environment. We show consistent performance during overload and throughput increases of up to 10 percent. Response time improves by up to a factor of 14, with only a 15 percent penalty to large jobs.

Variable latency on the Internet is a well-known problem for interactive applications. With the increase in interactive network games comes the increased importance of understanding the effects of latency on user performance. Classes of network games such as First Person Shooters (FPS) and Real Time Strategy (RTS) differ in their user interaction model and hence susceptibility to variable latency. While previous work has measured the effects of latency on FPS games, there has been no systematic investigation of the effects of latency on RTS games. In this work, we design and conduct user studies that measure the impact of latency on user performance in Warcraft III, a popular RTS game. As a foundation for the research, we separated typical Warcraft III user interactions into the basic components of explore, build and combat, and analyzed each individually. We nd modest statistical correlations with latency for exploration, but very weak correlations for building and combat. Overall, the effect of even very high latency, while noticeable to users, has a negligible effect on the outcome of the game. We attribute this somewhat surprising result to the nature of RTS game-play that clearly favors strategy over the real-time aspects.

This paper presents an architecture and algorithms for optimizing the performance of web services. For a given service, session-based admission control is combined with stage-wise request queuing, where the stages represent sub-tasks within sessions. The scheduling of requests is governed by generalized processor sharing. We present a performance model, relying on online estimation of parameters describing client-server interaction. A reward function corresponding to the service provider's objective is maximized using techniques for nonlinear optimization. In a case study, we model and optimize the resource sharing at a web server hosting an electronic store. The performance advantages of our approach are quantified numerically, and the robustness to parameter estimation errors is assessed by sensitivity analysis.

Latency on the Internet is a well-known problem for interactive applications. The growth in interactive network games brings an increased importance in understanding the effects of latency on user performance. Classes of network games such as First Person Shooters (FPS) and Real-Time Strategy (RTS) dier in their user interaction model and hence susceptibility to latency. While previous work has measured the effects of latency on FPS games, there has been no systematic investigation of the effects of latency on RTS games. In this work, we design and conduct user studies that measure the impact of latency on user performance on three of the most popular RTS games. As a foundation for the research, we separated typical RTS user interactions into the basic components of explore, build and combat, and analyzed each individually. We find modest statistical correlations between user performance and latency for exploration, but very weak correlations for building and combat. Overall, the effect of even very high latency, while noticeable to users, has a negligible effect on the outcome of the game. We attribute this somewhat surprising result to the nature of RTS game-play that clearly favors strategy over the real-time aspects.

This paper demonstrates that the dependability of generic, evolving J2EE applications can be enhanced through a combination of a few recovery-oriented techniques. Our goal is to reduce downtime by automatically and efficiently recovering from a broad class of transient software failures without having to modify applications. We describe here the integration of three new techniques into JBoss, an open-source J2EE application server. The resulting system is JAGR---JBoss with Application-Generic Recovery---a self-recovering execution platform.

Internet measurements show that a small number of large TCP flows are responsible for the largest amount of data transferred, whereas most of the TCP sessions are made up of few packets. Several authors have invoked this property to suggest the use of scheduling algorithms which favor short jobs, such as LAS (Least Attained Service), to differentiate between short and long TCP flows.

In this paper we describe the Medusa proxy, a tool for exploring user-perceived Web performance. The Medusa proxy is a non-caching forwarding proxy used in conjunction with a user's browser. The two key features that the Medusa proxy provides are (1) the ability to simultaneously mirror HTTP requests from the browser to different Web delivery systems and directly compare the results, and (2) the ability to transform requests, e.g., to transform Akamai URLs back into URLs to the customer's origin server.

Abstract — Energy and cooling costs of web server farms are among their main financial expenditures. This paper explores the benefits of dynamic voltage scaling (DVS) for power management in server farms. Unlike previous work, which addressed DVS on individual servers and on load-balanced server replicas, this paper addresses DVS in multi-stage service pipelines. Contemporary Web server installations typically adopt a three-tier architecture in which the first tier presents a Web interface, the second executes scripts that implement business logic, and the third serves database accesses. From a user’s perspective, only the end-to-end response across the entire pipeline is relevant. This paper presents a rigorous optimization methodology and an algorithm for minimizing the total energy expenditure of the multi-stage pipeline subject to soft end-to-end responsetime constraints. A distributed power management service is designed and evaluated on a real three-tier server prototype for coordinating DVS settings in a way that minimizes global energy consumption while meeting end-to-end delay constraints. The service is shown to consume as much as 30 % less energy compared to the default (Linux) energy saving policy.