An Investigation into the Application of Different Performance Prediction Methods to Distributed Enterprise Applications

Response time predictions for workload on new server architectures can enhance Service Level Agreement–based resource management. This paper evaluates two performance prediction methods using a distributed enterprise application benchmark. The historical method makes predictions by extrapolating from previously gathered performance data, while the layered queuing method makes predictions by solving layered queuing networks. The methods are evaluated in terms of: the systems that can be modelled; the metrics that can be predicted; the ease with which the models can be created and the level of expertise required; the overheads of recalibrating a model; and the delay when evaluating a prediction. The paper also investigates how a prediction-enhanced resource management algorithm can be tuned so as to compensate for predictive inaccuracy and balance the costs of SLA violations and server usage.     

Efficient lock-step synchronization in task-parallel languages

Many modern task-parallel languages allow the programmer to synchronize tasks using high-level constructs like barriers, clocks, and phasers. While these high-level synchronization primitives help the programmer express the program logic in a convenient manner, they also have their associated overheads. In this paper, we identify the sources of some of these overheads for task-parallel languages like X10 that support lock-step synchronization, and propose a mechanism to reduce these overheads. We first propose three desirable properties that an efficient runtime (for task-parallel languages like X10, HJ, Chapel, and so on) should satisfy, to minimize the overheads during lock-step synchronization. We use these properties to derive a scheme to called uClocks to improve the efficiency of X10 clocks; uClocks consists of an extension to X10 clocks and two related runtime optimizations. We prove that uClocks satisfies the proposed desirable properties. We have implemented uClocks for the X10 language+runtime and show that the resulting system leads to a geometric mean speedup of 5.36× on a 16-core Intel system and 11.39× on a 64-core AMD system, for benchmarks with a significant number of synchronization operations.     

“It was an artefact not the result”: A note on systems dynamic model development tools

Environmental modelling is done more and more by practising ecologists rather than computer scientists or mathematicians. This is because there is a broad spectrum of development tools available that allows graphical coding of complex models of dynamic systems and help to abstract from the mathematical issues of the modelled system and the related numerical problems for estimating solutions. In this contribution, we study how different modelling tools treat a test system, a highly non-linear predator–prey model, and how the numerical solutions vary. We can show that solutions (a) differ if different development tools are chosen but the same numerical procedure is selected; (b) depend on undocumented implementation details; (c) vary even for the same tool but for different versions; and (d) are generated but with no notifications on numerical problems even if these could be identified. We conclude that improved documentation of numeric methods used in the modelling software is essential to make sure that process based models formulated in terms of these modelling packages do not become “black box” models due to uncertainty in integration methods.     

Decision tree building on multi‐core using FastFlow

The whole computer hardware industry embraced the multi‐core. The extreme optimisation of sequential algorithms is then no longer sufficient to squeeze the real machine power, which can be only exploited via thread‐level parallelism. Decision tree algorithms exhibit natural concurrency that makes them suitable to be parallelised. This paper presents an in‐depth study of the parallelisation of an implementation of the C4.5 algorithm for multi‐core architectures. We characterise elapsed time lower bounds for the forms of parallelisations adopted and achieve close to optimal performance. Our implementation is based on the FastFlow parallel programming environment, and it requires minimal changes to the original sequential code. Copyright © 2013 John Wiley & Sons, Ltd.     

Automated Composition of Module Chains

We introduce a problem of module composition. Modules are seen as “black boxes” with input and output ports. A compatibility relation models which input ports can connected to which output ports. We are given a set of available modules and a target module. We want to connect available modules into a chain that implements the target module. Constraints may be given on how many copies of each module can or should appear in the solution chain. Costs may be given on modules or ports or connections, so that an optimal solution with respect to these costs is found. We derive an algorithm to solve the above problem automatically. The algorithm transforms the problem into a shortest-path problem in a graph.     

SAC—A Functional Array Language for Efficient Multi-threaded Execution

We give an in-depth introduction to the design of our functional array programming language SaC, the main aspects of its compilation into host machine code, and its parallelisation based on multi-threading. The language design of SaC aims at combining high-level, compositional array programming with fully automatic resource management for highly productive code development and maintenance. We outline the compilation process that maps SaC programs to computing machinery. Here, our focus is on optimisation techniques that aim at restructuring entire applications from nested compositions of general fine-grained operations into specialised coarse-grained operations. We present our implicit parallelisation technology for shared memory architectures based on multi-threading and discuss further optimisation opportunities on this level of code generation. Both optimisation and parallelisation rigorously exploit the absence of side-effects and the explicit data flow characteristic of a functional setting.     

Parallel DYNEMO: Meso-Scopic Traffic Flow Simulation on Large Networks

Traffic flow simulation is recently being applied to types of studies which require at the same time a large study area and models deeper than macroscopic assignments. Examples are investigations of the effects dynamic route guidance over large networks and the coupling with environmental models on large domains. The computational complexity can be reduced first by moving from microscopic to mesoscopic traffic flow models of the Payne-Cremer type. Where still more computing power is needed, parallelisation of the model may offer a solution. This paper describes a particular mesoscopic traffic flow model and its parallelisation undertaken in the ESPRIT project SIMTRAP. A new two step decomposition strategy is detailed along with its implementation using the message passing model PVM. The resulting speed-up is reported, drawing on both theoretical considerations and on measurements taken in the SIMTRAP demonstrator applications. The paper concludes with lessons learned from the experiments and directions for future work.     

有向图并行计算中的多目标剖分算法

在以离散网格为基础的某些数值模拟中,网格间的数据依赖关系可以抽象为有向图．如何剖分这些有向图成多个子图,将各子图对应的数值模拟任务映射到不同的处理机,是该类数值模拟并行计算的基础．剖分算法中,需要综合考虑连通性、并行度、负载平衡、通信开销四个目标．文章在传统有向图剖分算法的基础上,提出了一个权衡这四个目标的有向图多目标剖分区域分解算法．应用于二维非结构网格上的柱对称中子输运并行计算中,通量扫描并行算法在该区域剖分算法上获得的并行效率比原来的无向图区域剖分算法高50％以上．     

A Structured Synchronization and Communication Model Fitting Irregular Data Accesses

In this paper, we present a parallel programming and execution model based on alogicalordering of control flows. We show that it is possible to provide a unifying framework consisting of a synchronous programming model, thereby facilitating the mastery of programs, and an asynchronous execution model yielding efficient executions. Our approach is based on a SPMD and task parallel programming language, called –Chan. Communications take place through channels and rely on explicit send/receive instructions. In contrast to classical message passing models, synchronizations and communications are dissociated. We show that it is possible to perform a data-driven automatic translation of sequential and arbitrary DOACROSS loops into –Chan, by using nonmatching send/receive instructions. Our parallelization technique allows us to handle irregular control and leads to optimizations of communications in irregular computations.     

Online reputation systems: Design and strategic practices

This paper provides a comprehensive framework for evaluating the effects of feedback systems, and the potential problems with feedback systems, on seller incentives to provide high quality products under the case of asymmetric information. In particular, a baseline auction system without a feedback-based reputation system is first modeled and it is shown that the market for the sale of high quality products may not be sustainable. We then show how the institution of a feedback system can lead to sustainable market outcomes. Next, we demonstrate how three practices – the changing of identifications by dishonest sellers, shilling, and failure to leave feedback – can negate the usefulness of the feedback system in maintaining the market. Finally, we describe some actions that may be taken to overcome these problems. These results highlight the important role market managers play in ameliorating these market failure problems and ensuring the effective functioning of a market.     

A high performance engine for concurrent complex event processing

This paper describes the architecture, prototype implementation and performance analysis of a complex event processing engine that can scale up to very large numbers of concurrent events while keeping the requirements on system resources predictable and low. The main innovation of this approach is that each instantiated event pattern is handled by a dedicated Erlang process, instead of a single or shared operating system thread. This in turn, reduces the latency in processing the event processing as it avoids the overheads associated with resource contention. We demonstrate how this approach can achieve linear event processing times under high event loads, using modest computing resources. Copyright © 2013 John Wiley & Sons, Ltd.     

A unified approach to modelling the performance of concurrent systems

Quantitative design is crucial to ICT and it is therefore important to integrate performance modelling techniques into support environments that facilitate the correct construction of computer systems. We consider Performance Modelling Interchange Formats (PMIFs), which allow models to be specified in a uniform way and ported to a number of tools that solve them. We focus on extending the class of models describable in a PMIF that can be solved analytically – specifically, yielding a product-form solution for their equilibrium state probabilities. We use an extension of an established theorem, called the ‘reversed compound agent theorem’ (RCAT) as the basis of the analytical modelling tool into which the extended PMIF feeds models. We describe the RCAT methodology in practical terms, how it is integrated into an extended PMIF, and illustrate our methodology with three examples.     

Parallelisation of the Lagrangian atmospheric dispersion model NAME

The NAME  Atmospheric Dispersion Model is a Lagrangian particle model used by the Met Office to predict the propagation and spread of pollutants in the atmosphere. The model is routinely used in emergency response applications, where it is important to obtain results as quickly as possible. This requirement for a short runtime and the increase in core number of commonly available CPUs, such as the Intel Xeon series, has motivated the parallelisation of NAME  in the OpenMP  shared memory framework. In this work we describe the implementation of this parallelisation strategy in NAME  and discuss the performance of the model for different setups. Due to the independence of the model particles, the parallelisation of the main compute intensive loops is relatively straightforward. The random number generator for modelling sub-grid scale turbulent motion needs to be adapted to ensure that different particles use independent sets of random numbers. We find that on Intel Xeon X5680 CPUs the model shows very good strong scaling up to 12 cores in a realistic emergency response application for predicting the dispersion of volcanic ash in the North Atlantic airspace. We implemented a mechanism for asynchronous reading of meteorological data from disk and demonstrate how this can reduce the runtime if disk access plays a significant role in a model run. To explore the performance on different chip architectures we also ported the part of the code which is used for calculating the gamma dose from a cloud of radioactive particles to a graphics processing unit (GPU) using CUDA-C. We were able to demonstrate a significant speedup of around one order of magnitude relative to the serial CPU version.     

Super-human multi-talker speech recognition: A graphical modeling approach

We present a system that can separate and recognize the simultaneous speech of two people recorded in a single channel. Applied to the monaural speech separation and recognition challenge, the system out-performed all other participants – including human listeners – with an overall recognition error rate of 21.6%, compared to the human error rate of 22.3%. The system consists of a speaker recognizer, a model-based speech separation module, and a speech recognizer. For the separation models we explored a range of speech models that incorporate different levels of constraints on temporal dynamics to help infer the source speech signals. The system achieves its best performance when the model of temporal dynamics closely captures the grammatical constraints of the task. For inference, we compare a 2-D Viterbi algorithm and two loopy belief-propagation algorithms. We show how belief-propagation reduces the complexity of temporal inference from exponential to linear in the number of sources and the size of the language model. The best belief-propagation method results in nearly the same recognition error rate as exact inference.     

Subspace angles between ARMA models

We define a notion of subspace angles between two linear, autoregressive moving average, single-input–single-output models by considering the principal angles between subspaces that are derived from these models. We show how a recently defined metric for these models, which is based on their cepstra, relates to the subspace angles between the models.     

Measuring e-government readiness

We proposed a way of assessing readiness of a government organization to transform itself into a provider of fully integrated e-government services. We identified major components of e-government and discussed how it could evolve from a simple website into a fully integrated portal that delivers services to the public. Drawing upon strategic IS theory, we proposed three levels – strategic, system, and data – at which the governmental office’s readiness for e-government should be addressed. We empirically tested our model of e-government readiness with data collected from a U.S. municipal government. However, the results supported only a two-level model rather than the three we had theorized. This finding had important implications because it supported our contention that discrepancies can occur when e-government initiatives are at a relatively early stage and some critical e-government issues have not yet emerged.     

Data structures for the distributed iterative solution of non-conventional finite element models

A class of specialised data structures designed for the distributed solution of non-conventional finite element formulations, which are equally effective when used in conjunction with conventional formulations, is presented. We begin by briefly discussing how the non-conventional finite element formulations being developed within the structural analysis group at IST [Freitas JAT, Almeida JPM, Pereira EMBR. Non-conventional formulations for the finite element method. Comput Mech 1999;23(5–6):488–501] lead to systems of equations that appear to be naturally suited for parallel processing, but we also recognise that to take full advantage of the characteristics of these systems – large dimension, non-overlapping block structure and sparsity – it is necessary to use appropriate data structures. The approach presented, which references the logical subdivisions of the system matrices, was designed to fulfil these objectives. Examples of parallel performance and efficiency on an homogeneous distributed platform are presented.     

Understanding factors affecting trust in and satisfaction with mobile banking in Korea: A modified DeLone and McLean’s model perspective

As mobile technology has developed, mobile banking has become accepted as part of daily life. Although many studies have been conducted to assess users’ satisfaction with mobile applications, none has focused on the ways in which the three quality factors associated with mobile banking – system quality, information quality and interface design quality – affect consumers’ trust and satisfaction. Our proposed research model, based on DeLone and McLean’s model, assesses how these three external quality factors can impact satisfaction and trust. We collected 276 valid questionnaires from mobile banking customers, then analyzed them using structural equation modeling. Our results show that system quality and information quality significantly influence customers’ trust and satisfaction, and that interface design quality does not. We present herein implications and suggestions for further research.     

Stochastic programming and scenario generation within a simulation framework: An information systems perspective

Stochastic programming brings together models of optimum resource allocation and models of randomness to create a robust decision-making framework. The models of randomness with their finite, discrete realisations are called scenario generators. In this paper, we investigate the role of such a tool within the context of a combined information and decision support system. We explain how two well-developed modelling paradigms, decision models and simulation models can be combined to create “business analytics” which is based on ex-ante decision and ex-post evaluation. We also examine how these models can be integrated with data marts of analytic organisational data and decision data. Recent developments in on-line analytical processing (OLAP) tools and multidimensional data viewing are taken into consideration. We finally introduce illustrative examples of optimisation, simulation models and results analysis to explain our multifaceted view of modelling. In this paper, our main objective is to explain to the information systems (IS) community how advanced models and their software realisations can be integrated with advanced IS and DSS tools.