By Topic

Performance, Computing and Communications, 1998. IPCCC '98., IEEE International

Date 16-18 Feb. 1998

Filter Results

Displaying Results 1 - 25 of 60
  • 1998 IEEE International Performance, Computing and Communications Conference. Proceedings (Cat. No.98CH36191)

    Publication Year: 1998
    Save to Project icon | Request Permissions | PDF file iconPDF (493 KB)  
    Freely Available from IEEE
  • Author index

    Publication Year: 1998 , Page(s): 0_20
    Save to Project icon | Request Permissions | PDF file iconPDF (69 KB)  
    Freely Available from IEEE
  • Workload characterization using lightweight system call tracing and reexecution

    Publication Year: 1998 , Page(s): 260 - 266
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (672 KB)  

    This paper shows how system call traces can be obtained with minimal interference to the system being characterized, and used as realistic, repeatable workloads for experiments to evaluate operating system and file system designs and configuration alternatives. Our system call trace mechanism, called ULTra, captures a complete trace of each UNIX process's calls to the operating system. The performance impact is normally small, and it runs in user mode without special privileges. We show how the resulting traces can be used to drive full, repeatable reexecution of the captured behaviour, and present a case study which shows the usefulness and accuracy of the tool for predicting the impact of file system caching on a WWW server's performance View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Quality of service guarantees at end-to-end transport protocols

    Publication Year: 1998 , Page(s): 321 - 327
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (528 KB)  

    Emerging high-speed networks provide support for a wide range of applications with different traffic characteristics and quality of service (QoS) requirements. For example, real time applications require guaranteed services such as minimum bandwidth and bounded end-to-end delay. Network resources should be allocated in order to provide the guaranteed services required by applications. Networks employ traffic and congestion control mechanisms in order to allocate the needed network resources and enforce the allocation of these resources throughout the connection life. Moreover, QoS should also be guaranteed end-to-end at the transport protocol level. In this paper, we examine how QoS requirements for applications can be guaranteed at connection ends. In addition, we introduce a schedulability algorithm for applications that have a strict delay bound requirement. This algorithm will verify that packets delay bounds for these applications will not be violated View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Performance of distributed database application models using Java RMI

    Publication Year: 1998 , Page(s): 187 - 192
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (552 KB)  

    Java RMITM (remote method invocation) technology provides programmers with clean abstractions and simplified semantics for implementing parallel distributed programs on networks of heterogenous computing machines. This technology is being applied to many classes of applications in both the scientific/technical and business domains. From a performance perspective, is Java RMI technology a practical mode of expression for parallel distributed programs, and if so, what are the guidelines for architecting such programs? In this paper, we discuss the problem of expressing a parallel distributed program in Java and using Java RMI. To gauge the practicality of such an approach, we introduce a parametric benchmark based on artificially generated parallel distributed database programs. We then present an analysis of the observed performance of these programs on a network of workstations View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Congestion control in intelligent network

    Publication Year: 1998 , Page(s): 279 - 283
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (508 KB)  

    In the intelligent network (IN) service control points (SCPs) working as central servers carry out service requests coming from service switching points (SSPs). This paper develops an effective SCP overload control method and evaluates its characteristics. The characteristics are compared with other methods (window and call gapping) under various overload traffic patterns by using the network simulator that models SCP, SSPs, and their relationships in the IN View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The counter-based congestion avoidance scheme for ABR traffic in ATM networks

    Publication Year: 1998 , Page(s): 284 - 290
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (560 KB)  

    Congestion control plays an important role in the effective and stable operation of ATM networks. Many congestion control schemes for ABR traffic in ATM networks have been developed. However, conventional feedback schemes are very sensitive to network parameter settings and perform poorly when measured by throughput or queue length metrics. To solve these problems, a counter-based congestion avoidance scheme is proposed. The proposed scheme exhibits insensitivity to network parameter settings and has higher throughput and shorter queue length than previous schemes View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Estimation of error blocks in SDH networks

    Publication Year: 1998 , Page(s): 339 - 345
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (468 KB)  

    In this paper, we study bit interleaved parity (BIP) code performance based on the features of BIP code. We find the closed form by applying the binomial theorem to solve the undetected error probability when even parity is used for detecting transmission errors in SDH networks. We estimate BIP code performance and calculate the theoretical and detected values of block error rate according to the formulae recommended by ITU-T and amended by us. From a group of curves related to them, we study the numerical results. Our method provides a new choice for performance analysis of error blocks in SDH network, that is simpler than previous methods View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Data network analysis using NOVA

    Publication Year: 1998 , Page(s): 124 - 130
    Cited by:  Patents (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (808 KB)  

    A flexible, model-based tool is presented for use in the analysis of telecommunication networks. The tool was designed and implemented using object-oriented principles. Through the use of software models, which can be modified or added, NOVA allows the analysis of data networks with arbitrary characteristics and topologies. We discuss specific models for a network node and a link, which are the main components of a network topology. A sensitivity analysis of various model parameters was performed using a sample topology. In addition, NOVA was used to analyze the performance of the telecommunication network of southern China. We evaluated the network for various air traffic loads using fixed background data traffic. We investigated the average number of packets, average waiting time for a packet, average total time spent for a packet, and the utilization at a given city node in the network. By exploiting NOVA's modularity and flexibility, we were able to gain insight into various trade-offs involved in a telecommunication network design View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An overview and application of model reduction techniques in formal verification

    Publication Year: 1998 , Page(s): 165 - 171
    Cited by:  Papers (1)  |  Patents (3)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (616 KB)  

    Formal verification methods are becoming increasingly popular in the verification of digital systems and protocol checking of communication systems due to their inherent exhaustive nature. Model checking in particular has become quite extensively used in verifying the performance and functional specification of sequential logic. One characteristic of model checking is the state space explosion problem: as the number of state variables in the model under test increases, state space (and hence CPU/memory consumption of the model checker) increases exponentially. Whereas most previous papers on formal verification dealt largely with theoretical aspects, or were case studies and bug reports, the purpose of this paper is to provide an overview of state space reduction techniques, and then to illustrate several practical manual reduction techniques which have been employed successfully in the model checking of a node controller View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Generating dynamically scheduled memory address traces

    Publication Year: 1998 , Page(s): 245 - 250
    Cited by:  Papers (4)  |  Patents (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (664 KB)  

    It is common practice to study the performance of a computer memory hierarchy design via trace-driven simulation. The validity of such studies is dependent on the accuracy of the memory address trace that is used to drive the simulations. To study the performance of a memory hierarchy for a multiprogrammed computer system, the accuracy of the trace is defined by the length, in terms of the number of memory addresses, of the trace segments that comprise each process trace and the method used to interleave process trace segments. This paper reviews methods that have been used to generate memory address traces for use in simulating memory hierarchy designs. More importantly, it presents a new and more accurate method of generating these traces. Given a set of annotated uniprogram memory address traces (UMATs), which are associated with the processes that comprise the workload being studied, a specification of the design of the processor cycle time, memory hierarchy design, system call service times, and scheduling policy, this method dynamically generates a trace at simulation time by defining the correct trace segment interleaving via the correct process scheduling. One set of UMATs can be used to generate accurate traces for a variety of configurations of simulated systems View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Performance analysis of shared buffer ATM switch with different cell departure models

    Publication Year: 1998 , Page(s): 386 - 392
    Cited by:  Papers (2)  |  Patents (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (568 KB)  

    The performance analysis of an asynchronous transfer mode (ATM) switch involves modeling the input traffic source, the switching mechanism, and the cell departure process. A main issue which determines the overall accuracy in performance evaluation of ATM switch, is the use of an appropriate probabilistic model to describe the cell departure process. Until now, little work has been done in the literature on characterizing the cell departure process for studying the performance of an ATM switch. This paper compares and verifies the accuracy of several models, including a new one called the “Urn model”, which we propose. These models are put under test in a performance evaluation of a shared buffer ATM switch, by using a discrete-time Markov chain. The numerical results are compared to the simulation, and they show that the Urn model is a good compromise between accuracy and efficiency. This finding is significant because it helps to speed up the running of an analytical model of a large network while providing satisfactory accuracy View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • New network QoS measures for FEC-based audio applications on the Internet

    Publication Year: 1998 , Page(s): 355 - 362
    Cited by:  Papers (2)  |  Patents (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (532 KB)  

    New network QoS (quality of service) measures for interactive audio applications using FEC (forward error control) are proposed. Applications such as an Internet phone require both low data loss and a short delay for the underlying transport layer. The FEC-based error control has become popular as a way of meeting such requirements; new application data (or highly compressed data) are copied onto successive packets, so that random packet losses in networks can be concealed to some extent. From the viewpoint of FEC-based applications, actual QoS depends not only on the loss and delay of each packet, but also on the loss and delay of successive packets. Conventional network QoS measures such as loss rate and delay distribution, however, only focus on each packet. Therefore, the probability of long successive losses, for example, cannot be monitored, even though they strongly affect FEC-based application QoS. We propose a new concept named “loss window size” for measuring the QoS of successive packets. Definitions of loss and delay are generalized using this concept. These definitions take the FEC-based error concealment into account. Therefore, these measures enable more precise estimation of FEC-based application-level QoS than conventional measures. In order to show the effectiveness of the proposed measures, we have built an experimental monitoring system on working networks. The actual data show that network QoS may vary from time to time in terms of newly defined measures, even though QoS variation using conventional measures are not so apparent. We also model FEC-based application QoS, and show that application QoS and proposed network measures correspond well View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • ATM admission control based on measurements and reservations

    Publication Year: 1998 , Page(s): 298 - 304
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (712 KB)  

    We propose an innovative control architecture for ATM that allows users to reserve in advance complex connection patterns in the network. It offers control over the resources via schedules but alleviates the potential over-conservative nature of the resource allocation by coupling it with real-time measurement of actual use of the resources View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A practical building block for solving agreement problems in asynchronous distributed systems

    Publication Year: 1998 , Page(s): 25 - 31
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (772 KB)  

    Providing processes with the same view of a global state or allowing them to take consistent decisions, despite asynchrony and failure occurrences, are fundamental problems encountered in distributed systems. These problems are called agreement problems. Non blocking atomic commitment and definition of a single delivery order for broadcast messages are examples of such problems. We define a paradigm (called Single Global View) that encompasses various practical agreement problems. The interest of this paradigm lies in its practicability: each process starts with an initial value, and all these values are pieced together in such a way that, despite process crashes and asynchrony, all correct processes are delivered the same set of values (namely, the Single Global View). The power of this paradigm is the same as that of the consensus problem defined by theoreticians. Instantiations of the paradigm, which solve practical agreement problems, are given. A protocol implementing the paradigm is also presented View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Scalable multicast data distribution for different transport service classes

    Publication Year: 1998 , Page(s): 435 - 441
    Cited by:  Papers (3)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (668 KB)  

    Not all of the problems of large scale multicast can be addressed by automated repeat request (ARQ) based multicast protocols. Simulation studies have shown that independent packet loss among receivers is a worst case situation, where ARQ based protocols scale poorly and react with throughput degradation. Software based forward error correction (FEC) techniques are therefore a good solution and can be used to enhance the throughput efficiency by controlling the packet loss rate. A new method for implementing FEC at the transport layer is presented and performance measurements are shown. The difference between the restricted reliable multicast protocol (RRMP) and existing ARQ/FEC based multicast protocols is the flexibility and adaptability the RRMP provides by offering different transport service classes for different application needs. An application is able to choose the transport service, which best meets the requirements for efficiency, latency, delay, scalability etc. The design and implementation decisions for the FEC and ARQ based error recovery are outlined. Performance measurements demonstrate the suitability of FEC for real-time data transmissions View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Performance analysis of resource reservation strategies in broadband networks

    Publication Year: 1998 , Page(s): 307 - 313
    Cited by:  Patents (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (740 KB)  

    High speed networks are fast becoming an essential part of the information infrastructure. New problems in dynamic control of these networks arise due to the increased ratio of propagation delay to bit duration. Key network performance parameters (e.g., delay, cell loss, jitter) have to be observed while carrying various combinations of different types of traffic. This paper studies adaptive control strategies for dynamic resource allocation in broadband networks. Research goals include the development of adaptive control policies and dynamic resource scheduling algorithms which would efficiently and optimally allocate network resources (e.g., bandwidth and buffers), while maintaining the best possible quality of service for the different classes of calls. The use of reserved resources for each class of traffic in the dynamic allocation of network resources provides the required fairness and simplicity in implementing resource allocation control for high-speed networks while its adaptive control provides the network with an additional capability of adapting to the dynamics of the traffic load View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Performance optimisation evaluation of TCP/IP over wireless networks

    Publication Year: 1998 , Page(s): 395 - 401
    Cited by:  Papers (3)  |  Patents (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (540 KB)  

    The transmission control protocol (TCP) is a large extended transport protocol, mainly used in fixed networks. It works well in those networks in which the main cause of packet loss is congestion. Some mechanisms included in the TCP optimise their use in this kind of networks. Today, Internet users need access it over wireless links in order to get this service anywhere they are. However, the TCP does not work well in wireless environments where losses are due to high bit error rates (BER). We evaluate the end-to-end throughput of a TCP connection with a wireless link. First, we demonstrate the impact of high BER in the TCP performance. Then, we optimise the throughput of the TCP connection choosing the best MTU (maximum segment size) in each case, and evaluate other aspects of interests View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A simple hierarchical traffic model for VBR MPEG video

    Publication Year: 1998 , Page(s): 147 - 153
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (612 KB)  

    A hierarchical model is derived for MPEG video. This model is developed so as to represent the bit-rate variations of MPEG video at two different time scales/levels: scene level and frame level. Various evaluations show that the proposed model not only matches very good the statistical characteristics with the real MPEG video but also predicts accurately the network performance of an ATM multiplexer View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Delta service in a mobile file system client cache

    Publication Year: 1998 , Page(s): 222 - 228
    Cited by:  Patents (6)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (644 KB)  

    This paper evaluates a delta service extension to a mobile file system cache developed in order to minimize the amount of data transferred over wireless communications links. Network bandwidth stands as one of the resource limitations impacting the design of mobile computer applications. At the mobile file system service level, caching provides resource conservation in distributed applications. This paper proposes a delta service to enhance caching services characteristic of mobile computer file systems. Well established mechanisms for sequence comparison and software configuration management, file deltas have applicability to mobile computer and distributed file system caching environments. Study of the delta service uses trace-driven simulation methodology incorporating traces obtained in a real world distributed environment. A delta service model will extend the mobile computer client cache model of various cache sizes in order to gauge the bandwidth savings on the link obtained by the delta service View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Bounds-based loop performance analysis: application to validation and tuning

    Publication Year: 1998 , Page(s): 178 - 184
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (596 KB)  

    We consider the floating point microarchitecture support in high-end RISC superscalar processors. We propose a simple, yet effective bounds model to deduce the “bestcase” loop performance limits for these processors. We compare these bounds to simulation-based (and where available, hardware-based) performance measurements for actual compiler-generated code sequences. From this study, we identify loop tuning opportunities to bridge the gap between “best-case” and “actual” performance in a post-silicon setting. Some of the results of such analysis point to fundamental hardware performance bugs which may be removed through feasible microarchitectural changes. More frequently, the analysis is useful for suggesting compiler enhancements. The analysis methods described have been used in actual high-end processor development projects within our company. We report our experimental results in the context of a set of application-based loop test cases, designed to stress various resource limits in the core (infinite cache) microarchitecture View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A fault-tolerant distributed sorting algorithm in tree networks

    Publication Year: 1998 , Page(s): 37 - 43
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (684 KB)  

    The paper presents a distributed sorting algorithm for a network with a tree topology. The distributed sorting problem can be informally described as follows: nodes cooperate to reach a global configuration where every node, depending on its identifier, is assigned a specific final value taken from a set of input values distributed across all nodes; the input values may change in time. In our solution, the system reaches its final configuration in a finite time after the input values are stable and the faults cease. The fault tolerance and adaption to changing input are achieved using Dijkstra's paradigm of self stabilization. Our solution is based on continuous broadcast with acknowledgment along the tree network to achieve synchronization among processes in the system; it has O(nh) time complexity and only O(log(n)deg) memory requirement where deg is the degree of the tree and h is the height of the tree View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Modeling the performance of general purpose instruction level parallel architectures in image processing

    Publication Year: 1998 , Page(s): 267 - 273
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (588 KB)  

    RISC Instruction Level Parallel systems are today the most commonly used high performance computing platform. On such systems, Image Processing and Pattern Recognition (IPPR) tasks, if not thoroughly optimized to fit each architecture, exhibit a performance level up to one order of magnitude lower than expected. In this paper we identify the sources of such behavior and we model them defining a set of indices to measure their influence. Our model allows planning program optimizations, assessing the results of such optimizations as well as evaluating the efficiency of the CPUs architectural solutions in IPPR tasks. A case study using a combination of a specific IPPR task and a RISC workstation is used to demonstrate these capabilities. We analyze the sources of inefficiency of the task, we plan some source level program optimizations, namely data type optimization and loop unrolling, and we assess the impact of these transformations on the task performance. The results of our study allow us to obtain an eight times performance improvement and to conclude that, in low-medium level IPPR tasks, it is more difficult to efficiently exploit superscalarity than pipelining View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Channel reification: a reflective model for distributed computation

    Publication Year: 1998 , Page(s): 32 - 36
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (528 KB)  

    The paper presents a new reflective model, called Channel Reification, which can be used in distributed computations to overcome difficulties experienced by other models in the literature when monitoring communication among objects. The channel is an extension of the message reification model. A channel is a communication manager incarnating successive messages exchanges by two objects: its application ranges between those of message reification and those of meta object model. After a brief review of existing reflective models and how reflections can be used in distributed systems, channel reification is presented and compared to the widely used meta object model. Applications of channel reification to protocol implementation, and to fault tolerant object systems are shown. Future extensions to this model are also summarized View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • PLAN: a pipeline-access local ATM network protocol for multimedia applications

    Publication Year: 1998 , Page(s): 379 - 385
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (676 KB)  

    This paper presents a shared media ATM LAN architecture known as pipelined-access local ATM network (PLAN) suitable for both LAN and MAN applications. It is a reservation type bi-directional broadcast bus system protocol in which stations are connected at predefined points on the bus and all transmissions are done sequentially from one end to the other in a pipelined manner. It is designed to (i) support multimedia traffic, (ii) dynamically allocate bandwidth to various types of traffic and provide fair media access to stations, (iii) provide throughput close to 100%, (iv) operate at high bit rates, and, (v) use packets of ATM cell size and format for interoperability with the wide-area ATM networks. We present the architecture, analysis and performance of the proposed PLAN protocol under various multimedia traffic conditions. The results show that its throughput approaches 100%, is independent of traffic composition, and increases slowly with bit rate. It decreases only slightly as the bus length and/or number of stations increases. For example, the throughput decreases only by 0.17% per kilometer as the bus length increases. Based on the results obtained, the proposed PLAN protocol is seen to be well suited for high speed multimedia ATM LAN and MAN applications View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.