Catálogo de publicaciones - libros

Current implementations of Grid infrastructures provide frameworks which aim at achieve on-demand computing. In such a scenario, contribution and use of resources will be governed by business models. The challenge is to provide multi-level performance information which enables the participation of the different actors in such a system. In this paper we describe the performance measuring framework developed for Grid Market Middleware, a middleware which supports economic-model based selection of service-oriented Grid applications. This middleware is a distributed infrastructure, which we have implemented for providing a market of services and resources to be assigned to Grid applications. The objectives of the performance measuring framework is first to assess the behaviour of the middleware and the used economic models in a deployed system, and secondly allow the provision of metrics for the components of the middleware itself. We describe the design of the performance measuring framework, its implementation and show its capability and usefulness for our objectives by experiments.

In this paper we propose a new efficient iterative scheme for solving closed queueing networks with phase-type service time distributions. The method is especially efficient and accurate in case of large numbers of nodes and large customer populations. We present the method, put it in perspective, and validate it through a large number of test scenarios. In most cases, the method provides accuracies within 5% relative error (in comparison to discrete-event simulation).

A rather complex task in the performance analysis of software architectures has always been the interpretation of the analysis results and the generation of feedback that may help developers to improve their architecture with alternative ”better performing” solutions. This is due, on one side, to the fact that performance analysis results may be rather complex to interpret (e.g., they are often collections of different indices) and, on the other side, to the problem of coupling the ”right” architectural alternatives to results, that are the alternatives that allow to improve the performance by resolving critical issues in the architecture. In this paper we propose a framework to interpret the performance analysis results and to propose alternatives to developers that improve their architectural designs. The interpretation of results is based on the ability to automatically recognize performance anti-patterns in the software architecture. The whole process of result interpretation and generation of architectural alternatives is supported by a tool based on the Layered Queueing Network notation.

A system consisting of several servers, where demands of different types arrive in bursts, is examined. The servers can be dynamically reallocated to deal with the different requests, but these switches take time and incur a cost. The problem is to find the optimal dynamic allocation policy. To this end a Markov decision process is solved, using two different techniques. The effects of different solution methods and modeling decisions on the resulting solution are examined.

Current computer systems and communication networks tend to be highly complex, and they typically hide their internal structure from their users. Thus, for selected aspects of capacity planning, overload control and related applications, it is useful to have a method allowing one to find good and relatively simple approximations for the observed system behavior. This paper investigates one such approach where we attempt to represent the latter by adequately selecting the parameters of a set of queueing models. We identify a limited number of queueing models that we use as Building Blocks in our procedure. The selected Building Blocks allow us to accurately approximate the measured behavior of a range of different systems. We propose an approach for selecting and combining suitable Building Blocks, as well as for their calibration. We are able to successfully validate our methodology for a number of case studies. Finally, we discuss the potential and the limitations of the proposed approach.

We show how to combine censoring technique for Markov chain and strong stochastic comparison to obtain bounds on rewards and the first passage time. We present the main ideas of the method, the algorithms and their proofs. We obtain a substantial reduction of the state space due to the censoring technique. We also present some numerical results to illustrate the effectiveness of the method.

Message-oriented middleware (MOM) is at the core of a vast number of financial services and telco applications, and is gaining increasing traction in other industries, such as manufacturing, transportation, health-care and supply chain management. There is a strong interest in the end user and analyst communities for a standardized benchmark suite for evaluating the performance and scalability of MOM. In this paper, we present a workload characterization of the SPECjms2007 benchmark which is the world’s first industry-standard benchmark specialized for MOM. In addition to providing standard workload and metrics for MOM performance, the benchmark provides a flexible performance analysis framework that allows users to customize the workload according to their requirements. The workload characterization presented in this paper serves two purposes i) to help users understand the internal components of the SPECjms2007 workload and the way they are scaled, ii) to show how the workload can be customized to exercise and evaluate selected aspects of MOM performance. We discuss how the various features supported by the benchmark can be exploited for in-depth performance analysis of MOM infrastructures.

In software systems, individual components interact not only through explicit function invocations, but also through implicit resource sharing. The use of shared resources significantly influences the duration of the invoked functions. For resources that are heavily shared, capturing this influence can lead to performance models that have a large number of elements and a large number of dependencies. We introduce an approach that can model resource sharing separately from function invocations, keeping the performance model reasonably simple while still describing many of the effects of resource sharing on the duration of function invocations. The approach has been tested on the CoCoME component application modeling example.

In the last couple of years, hard-disk technology has experienced an unjustified progressive boost of the built-in cache size, affecting both the power consumption and the reliability of stored data. Large built-in caches offer limited benefits in terms of performance with respect to the smaller ones. Moreover, they need to be kept in write-through mode to preserve data in case of a power failure in mission-critical systems. This implies severe repercussions on the disk write performance, due to the role of the built-in cache itself, mainly acting as a write scheduler, rather than just a mere I/O buffer, as its ever increasing size would suggest.

In this scenario, an exact hard-disk characterization can provide the upper layers enough information to compensate the performance loss produced by the write-through policy. File-systems and device-drivers can in fact obviate most of these issues via proper data layouts, depending on a detailed knowledge of the hard-disks geometry.

This paper introduces the chunk skew layout, a novel data layout strategy targeted to improve the performance of commodity hard-disks in mission-critical systems. For this purpose we also analyze the differences in terms of geometry and performance in a batch of identical commodity hard-disks, discovering an unexpected and more complex scenario where most of the assumptions made so far on hard-disk technology do not hold anymore.

In mobile ad hoc networks (MANETs), broadcasting plays a fundamental role, diffusing a message from a given source node to all the other nodes in the network. Flooding is the simplest and commonly used mechanism for broadcasting in MANETs, where each node retransmits every uniquely received message exactly once. Despite its simplicity, it however generates redundant rebroadcast messages which results in high contention and collision in the network, a phenomenon referred to as . Pure probabilistic approaches have been proposed to mitigate this problem inherent with flooding, where mobile nodes rebroadcast a message with a probability which can be fixed or computed based on the local density. However, these approaches reduce the number of rebroadcasts at the expense of reachability. On the other hand, counter-based approaches inhibit a node from broadcasting a packet based on the number of copies of the broadcast packet received by the node within a random access delay time. These schemes achieve better throughput and reachability, but suffer from relatively longer delay. In this paper, we propose an efficient broadcasting scheme that combines the advantages of pure probabilistic and counter-based schemes to yield a significant performance improvement. Simulation results reveal that the new scheme achieves superior performance in terms of saved-rebroadcast, reachability and latency.