The underutilization of disk parallelism and file cache buffers by traditional file systems induces I/O stall time that degrades the performance of modern microprocessor-based systems. In this paper, we present aggressive mechanisms that tailor file system resource management to the needs of I/O-intensive applications. In particular, we show how to use application-disclosed access patterns (hints) to expose and exploit I/O parallelism and to allocate dynamically file buffers among three competing demands: prefetching hinted blocks, caching hinted blocks for reuse, and caching recently used data for unhinted accesses. Our approach estimates the impact of alternative buffer allocations on application execution time and applies a cost-benefit analysis to allocate buffers where they will have the greatest impact. We implemented informed prefetching and caching in DEC’s OSF/1 operating system and measured its performance on a 150 MHz Alpha equipped with 15 disks running a range of applications including text search, 3D scientific visualization, relational database queries, speech recognition, and computational chemistry. Informed prefetching reduces the execution time of the first four of these applications by 20 % to 87%. Informed caching reduces the execution time of the fifth application by up to 30%.

Sequentiality of requested blocks on disks, or their spatial locality, is critical to the performance of disks, where the throughput of accesses to sequentially placed disk blocks can be an order of magnitude higher than that of accesses to randomly placed blocks. Unfortunately, spatial locality of cached blocks is largely ignored and only temporal locality is considered in system buffer cache management. Thus, disk performance for workloads without dominant sequential accesses can be seriously degraded. To address this problem, we propose a scheme called DULO (DUal LOcality), which exploits both temporal and spatial locality in buffer cache management. Leveraging the filtering effect of the buffer cache, DULO can influence the I/O request stream by making the requests passed to disk more sequential, significantly increasing the effectiveness of I/O scheduling and prefetching for disk performance improvements. DULO has been extensively evaluated by both tracedriven simulations and a prototype implementation in Linux 2.6.11. In the simulations and system measurements, various application workloads have been tested, including Web Server, TPC benchmarks, and scientific programs. Our experiments show that DULO can significantly increase system throughput and reduce program execution times. 1

It is well known that IEEE 802.11 provides a physical layer multi-rate capability, and hence MAC layer mechanisms are needed to exploit this capability. Several solutions have been proposed to achieve this goal. However, these solutions only consider how to exploit good channel quality for the direct link between the sender and the receiver. Since IEEE 802.11 supports multiple transmission rates in response to different channel conditions, data packets may be delivered faster through a relay node than through the direct link if the direct link has low quality and low rate. In this paper, we propose a novel MAC layer relay-enabled distributed coordination function (DCF) protocol, called rDCF, to further exploit the physical layer multi-rate capability. We design a protocol to assist the sender, the relay node and the receiver to reach an agreement on which data rate to use and whether to transmit the data through a relay node. Considering various issues such as bandwidth utilization and channel errors, we propose techniques to further improve the performance of rDCF. Simulation results show that rDCF can significantly improve the system performance when the channel quality of the direct link is poor.

Current disk prefetch policies in major operating systems track access patterns at the level of the file abstraction. While this is useful for exploiting application-level access patterns, file-level prefetching cannot realize the full performance improvements achievable by prefetching. There are two reasons for this. First, certain prefetch opportunities can only be detected by knowing the data layout on disk, such as the contiguous layout of file metadata or data from multiple files. Second, non-sequential access of disk data (requiring disk head movement) is much slower than sequential access, and the penalty for mis-prefetching a ‘random ’ block, relative to that of a sequential block, is correspondingly more costly. To overcome the inherent limitations of prefetching at the logical file level, we propose to perform prefetching directly at the level of disk layout, and in a portable way. Our technique, called DiskSeen, is intended to be supplementary to, and to work synergistically with, filelevel prefetch policies, if present. DiskSeen tracks the locations and access times of disk blocks, and based on analysis of their temporal and spatial relationships, seeks to improve the sequentiality of disk accesses and overall prefetching performance. Our implementation of the DiskSeen scheme in the Linux 2.6 kernel shows that it can significantly improve the effectiveness of prefetching, reducing execution times by 20%-53 % for micro-benchmarks and real applications such as grep, CVS, and TPC-H. 1

Abstract — Due to the broadcast nature of wireless signals, a wireless transmission intended for a particular destination station can be overheard by other neighboring stations. A focus of recent research activities in cooperative communications is to achieve spatial diversity gains by requiring these neighboring stations to retransmit the overheard information to the final destination. In this paper we demonstrate that such cooperation among stations in a wireless LAN (WLAN) can achieve both higher throughput and lower interference. We present the design for a medium access control protocol called CoopMAC, in which high data rate stations assist low data rate stations in their transmission by forwarding their traffic. In our proposed protocol, using the overheard transmissions, each low data rate node maintains a table, called a CoopTable, of potential helper nodes that can assist in its transmissions. During transmission, each low data rate node selects either direct transmission or transmission through a helper node in order to minimize the total transmission time. Using analysis, simulation and testbed experimentation, we quantify the increase in the total network throughput, and the reduction in delay, if such cooperative transmissions are utilized. The CoopMAC protocol is simple and backward compatible with the legacy 802.11 system. In this paper, we also demonstrate a reduction in the signal-to-interference ratio in a dense deployment of 802.11 access points, which in some cases is a more important consequence of cooperation. Index Terms — IEEE 802.11, medium access control, protocol design and analysis, multi-rate, rate adaptation, cooperative networking, cross-layer design, bridges.

Abstract. The throughput capacity of WLANs can be improved by a carefully designed relay infrastructure. In this work, we propose an optimization formulation based on Lagrangian relaxation and a subgradient algorithm to compute the best placement of a fixed number of relay nodes (RNs) in a WLAN. We apply this optimization framework to a multi-rate WLAN based on the IEEE 802.11g standard under Rayleigh fading. We then study the expected throughput capacity of a WLAN with relay infrastructure and investigate how the optimal placement of RNs is affected by the number of RNs, path-loss characteristics, and the traffic pattern. Our numerical results show that, in some network scenarios, more than 120 % performance gain can be achieved when RNs are strategically installed in the network. Furthermore, we also show that for a wide range of system parameters, optimally placed RNs can significantly increase the network throughput capacity over random placement. Key words: WLAN, immobile relays, throughput capacity, optimal placement 1

With the proliferation of wireless local area network (WLAN) technologies, wireless Internet access via public hotspots will become a necessity in the near future. In outdoor areas where the installation of a large number of wired access points is practically or economically infeasible, mobile users located at the edge of the network communicate with the access point at a very low rate and in turn waste network resources. In this work, we promote the use of tetherless relay points (TRPs) to improve the throughput of a WLAN in such environments. We first provide a high level description on how to integrate TRPs in a multi-rate WLAN architecture. We then propose an integer-programming optimization formulation and an iterative approach to compute the best placement of a fixed number of TRPs. Finally, we show in numerical analysis, through a case study based on relay-enabled rate adaptation and IEEE 802.11-like multi-rate physical model with Rayleigh fading, that for a wide range of system parameters, significant performance gain can be achieved when TRPs are strategically installed in the network.

Abstract — We analyze the performance of the Distributed Coordination Function (DCF) for 802.11 WLANs. We consider a fixed number of contending stations within radio proximity, and we investigate the important case in which stations operate under non-saturated conditions. We assume that the MAC queues of wireless stations receive from the upper layers a stationary arrival process of packets. We identify the fundamental problems that arise in building an analytical model of the system, and we propose different approaches to overcome these difficulties. Finally, we apply our modelling technique to study several important issues in 802.11 networks, such as the impact of bursty traffic and the system performance in a multirate environment. The accuracy of the analytical results is verified by simulation with ns-2. Index Terms — Stochastic processes/Queueing theory I.

Abstract. In this paper, we present an analytic model for evaluating the queueing delays at nodes using the IEEE 802.11 Point Coordination Function (PCF) MAC for real time, delay sensitive traffic. We develop a queueing model for each node to obtain closed form expressions for the expected delay which accounts for arbitrary packet sizes, polling rates, channel rates and the order in which the nodes are polled. Our analytical results are verified through simulations.

. We have investigated the construction of a parallel computer using IEEE 1355 high-throughput low-latency DS link networks and high-performance commodity processors running a standard operating system. In this context a DS Network Interface Controller (DSNIC) has been developed. The board's hardware, controlled by FPGA firmware, together with host software, provides a CSP based message passing interface between standard OS processes. This paper describes how the design and realisation of the DSNIC reflect our aim: low-latency high-throughput inter-process communication. We show the benchmark results, their analysis, and suggest further performance gains that might be possible. 1 Introduction CERN is currently building a new accelerator to allow the study of collisions between elementary particles at very high energies. A major problem will be the huge amounts of data produced during a collision combined, with the frequency at which collisions occur. For example, the ATLAS[2] experime...