Most operating systems use interface interrupts to schedule network tasks. Interrupt-driven systems can provide low overhead and good latency at low of-fered load, but degrade significantly at higher arrival rates unless care is taken to prevent several pathologies. These are various forms of receive livelock, in which the system spends all its time processing interrupts, to the exclusion of other neces-sary tasks. Under extreme conditions, no packets are delivered to the user application or the output of the system. To avoid livelock and related problems, an operat-ing system must schedule network interrupt handling as carefully as it schedules process execution. We modified an interrupt-driven networking implemen-tation to do so; this eliminates receive livelock without degrading other aspects of system performance. We present measurements demonstrating the success of our approach. 1.

The explosive growth of the Internet, the widespread use of WWW-related applications, and the increased reliance on client-server architectures places interesting new demands on network servers. In particular, the operating system running on such systems needs to manage the machine’s resources in a manner that maximizes and maintains throughput under conditions of high load. We propose and evaluate a new network subsystem architecture that provides improved fairness, stability, and increased throughput under high network load. The architecture is hardware independent and does not degrade network latency or bandwidth under normal load conditions.

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.

mogul @ wrl.dec.com The Internet has experienced exponential growth in the use of the World-Wide Web, and rapid growth in the use of other Internet services such as VSENET news and electronic mail. These applications qualitatively differ from other network applications in the stresses they impose on busy server systems. Unlike traditional distributed systems, Internet servers must cope with huge user communities, short interactions, and long network latencies. Such servers require different kinds of operating system features to manage their resources effectively. 1

The World Wide Web and its related applications place substantial performance demands on network servers. The ability to measure the effect 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 in a test-bed environment. Unfortunately, the state-of-the-art approach for benchmarking Web servers is unable to generate client request rates that exceed the capacity of the server being tested, even for short periods of time. Moreover, it fails to model important characteristics of the wide area networks on which most servers are deployed (e.g. delay and packet loss). 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. Our method also mod...

Indoor environments present opportunities for a rich set of location-aware applications such as navigation tools for humans and robots, interactive virtual games, resource discovery, asset tracking, location-aware sensor networking etc. Typical indoor applications require better accuracy than what current outdoor location systems provide. Outdoor location technologies such as GPS have poor indoor performance because of the harsh nature of indoor environments. Further, typical indoor applications require different types of location information such as physical space, position and orientation. This dissertation describes the design and implementation of the Cricket indoor location system that provides accurate location in the form of user space, position and orientation to mobile and sensor network applications. Cricket consists of location beacons that are attached to the ceiling of a building, and receivers, called listeners, attached to devices that need location. Each beacon periodically transmits its location information in an RF message. At the same time,

this paper is largely motivated by this recent work by Chandra and Mosberger [6]. We believe that their work demonstrates that even simple server designs exhibit a wide range of variation in performance that is not well understood. In this paper we attempt to characterize the behaviour of some of these design options and use these results to gain insight into some of the issues affecting server designs in general

A QoS-aware real-time operating system must schedule multiple tasks which have dierent timing constraints and which access various resources including the CPU, disk and network. These resources, however, are not independent of one another. For example, resources like network bandwidth and disk bandwidth are available on a single node but must be managed by their host OS on the CPU by means of interrupt handlers, device drivers, le-systems and/or protocol services. Hence, in order to obtain guaranteed completion times, an application must therefore obtain both usermode time on the CPU along with sucient OS-level time for the network and disk subsystems. In this paper, we investigate the co-scheduling of CPU cycles and network bandwidth. Speci cally, we study the problem of obtaining pre-speci ed network bandwidth received by applications from the external network. Our solution endows (1)direct control over the ow of network packets into the system based on the requirements of speci c applications, (2)guaranteed and enforced processing time for the received packets, (3)precise accounting of those processing times, and (4)elimination of scheduling anomalies. We describe and evaluate our system design and implementation in Linux/RK, a QoS-aware real-time version of Linux. We also compare this approach with a commercial implementation we did in TimeSys Linux.

Abstract not found