Johannes Lüthi

The resource sharing system: dynamic federate mapping for HLA-based distributed simulation

Networks of workstations have become a popular architecture for distributed simulation due to their high availability as opposed to specialized multiprocessor computers. Networks of workstations are also a well-suited framework for distributed simulation systems based on the High Level Architecture (HLA). However using workstations in a distributed simulation system may eventually affect the availability of computing resources for the users who need their computers as working tools. Thus, for coarse grained distributed simulation it may be desirable to let the users control to what extent their workstations should participate in a distributed simulation. The authors present a resource sharing system (RSS) that provides a client user interface on each potentially participating workstation. With the RSS clients, users of workstations can control the availability of their computer for the HLA simulation federation. An RSS manager keeps track of available computing resources and balances the participating HLA federates among the available workstations.

Towards Time-Parallel Road Traffic Simulation

As an alternative to spatial parallelization of simulation models, time-parallel simulation offers the potential for massive parallelism with a high level of independence between the parallel processes. Unfortunately, due to inherent problems, the applicability of time-parallel simulation is restricted. Therefore, it has been proposed recently, to use approximation with time-parallel simulation in order to facilitate its application and to extend the class of models suitable for time-parallel simulation. As a proof-of-concept, this work shows how approximate temporal parallelization can be applied to the simulation of road traffic. Traffic simulation is used extensively in transportation research for various purposes, e.g. analysis of traffic phenomena, traffic forecast, and optimization of traffic flow. Depending on the level of fidelity, a traffic model exhibits a state space of moderate to high complexity. This paper is intended to discuss the basic properties of time-parallel traffic simulation and to examine its feasibility. Experiments with a sequential microscopic traffic simulator, that emulates important aspects of a corresponding time-parallel simulator, suggest this feasibility.

Mean value analysis for computer systems with variabilities in workload

When evaluating the performance of computer systems, often uncertainties or variabilities in service demands may be observed. Applying well known mean valve analysis (MVA) for single- or multiclass queueing network models of such systems is inappropriate and ineffective, because these models fail to represent variations within a class. This paper proposes to use histograms for characterizing model parameters that are associated with uncertainty or variability and presents an adaptation of the single class MVA algorithm, which traditionally accepts single (mean) values for service demands, so that one or more input parameters can be specified as a histogram. The adapted algorithm generates a histogram output for the performance measures, thus providing a more detailed information (e.g. percentile values) than the mean valves obtained from conventional MVA. The proposed technique is demonstrated on selected examples in different problem domains. It is shown, that the computational complexity is reasonable given that the number of parameters specified as histograms is not too high. Although the algorithm produces accurate results in many situations inaccuracies have been observed for certain systems. A technique called interval splitting that can be used for controlling such inaccuracies is described.

Mean value analysis for queueing network models with intervals as input parameters

Mean value analysis (MVA) is a well-known solution technique for separable closed queueing networks used in performance modeling of computer and communication systems. In many cases, like for sensitivity analysis or with inaccurate model input parameters, intervals are more appropriate as model inputs than single values. This paper presents a version of the MVA algorithm for separable closed queueing networks with one customer class consisting of load-independent queueing centers as well as delay devices, which accepts both single values and intervals as input parameters in an arbitrary combination. Monotonicity of the model outputs with respect to all input parameters is proved and these monotonicity properties are used to construct a low cost interval-version of the MVA algorithm providing exact output intervals as results. Thus, dependency problems commonly arising with the interval evaluation of arithmetic expressions are avoided without significant increase in computation costs. Additionally, asymptotic results for the performance measure intervals obtained through interval-based analysis and corresponding bottleneck analysis are presented.

Performance bounds for distributed systems with workload variabilities and uncertainties

Abstract Bounding techniques for queuing network models used to analyze the performance of parallel and distributed computer systems accept single values as model inputs. Uncertainties or variabilities in service demands may exist in many types of systems. Using models with a single aggregate mean value for each parameter for such systems can lead to inaccurate or even incorrect results. This paper proposes to use histograms for characterizing model parameters that are associated with uncertainty and/or variability. The adaptation of the well-known asymptotic bounds as well as balanced job bounds for single class queuing networks to histogram parameters is presented in the paper.

Interval matrices for the bottleneck analysis of queueing network models with histogram-based parameters

Bottleneck analysis using queueing network models is an important technique for the performance analysis and capacity planning of computer and communication systems. Conventional single class as well as multiclass queueing network models use single mean values as input parameters. However uncertainties and variabilities in service demands may exist in many models. This paper proposes to use extended histograms for characterizing model parameters that are associated with workload uncertainty and/or variability. Because with histogram-based parameters, system bottlenecks need not be unique, methods are presented which produce interval-based bottleneck identification matrices. Additionally, interval matrices for the approximation of potential effects of service demand modifications are presented. With the proposed interval matrix approach, associated input parameter variabilities and uncertainties are also represented in the model output. Thus, model uncertainties are not hidden but an overview of the potential model behavior is provided to the analyst.Bottleneck analysis using queueing network models is an important technique for the performance analysis and capacity planning of computer and communication systems. Conventional single class as well as multiclass queueing network models use single mean values as input parameters. However uncertainties and variabilities in service demands may exist in many models. This paper proposes to use extended histograms for characterizing model parameters that are associated with workload uncertainty and/or variability. Because with histogram-based parameters, system bottlenecks need not be unique, methods are presented which produce interval-based bottleneck identification matrices. Additionally, interval matrices for the approximation of potential effects of service demand modifications are presented. With the proposed interval matrix approach, associated input parameter variabilities and uncertainties are also represented in the model output. Thus, model uncertainties are not hidden but an overview of the potential model behavior is provided to the analyst.

Interval parameters for capturing uncertainties in an EJB performance model

Exact as well as approximate analytical solutions for quantitative performance models of computer systems are usually obtained by performing a series of arithmetical operations on the input parameters of the model. However, especially during early phases of system design and implementation, not all the parameter values are usually known exactly. In related research contributions, intervals have been proposed as a means to capture parameter uncertainties. Furthermore, methods to adapt existing solution algorithms to parameter intervals have been discussed. In this paper we present the adaptation of an existing performance model to parameter intervals. The approximate solution of a queueing network modelling an Enterprise JavaBeans server implementation is adapted to interval arithmetic in order to represent the uncertainty in some of the parameters of the model. A new interval splitting method is applied to obtain reasonable tight performance measure intervals. Monotonicity properties of intermediate computation results are exploited to achieve a more efficient interval solution. In addition, parts of the original solution algorithm are modified to increase the efficiency of the corresponding interval arithmetical solution.

Characterization and Analysis of Software and Computer Systems with Uncertainties and Variabilities

Conventional solution techniques for analytic performance models of computer and telecommunication systems use single values as inputs. Uncertainties or variabilities in model parameters may exist in many types of systems. Using models with a single aggregated mean value for each parameter for such systems can produce inappropriate and misleading results. This chapter presents intervals and extended histograms for characterizing system parameters that are associated with uncertainty and variability. Adaptation of existing analytic performance evaluation methods to this interval-based parameter characterization is described. The application of this approach is illustrated with two examples: a hierarchical model of a multicomputer system and a queueing network model of an EJB server implementation.

FT-RSS: A Flexible Framework for Fault Tolerant HLA Federations

The absence of fault tolerance mechanisms is a significant deficit of most current distributed simulation in general and of simulation systems based on the high level architecture (HLA) in particular. Depending on failure assumptions, dependability needs, and requirements of the simulation application, a choice of different mechanisms for error detection and error processing may be applied. In this paper we propose a framework for the configuration and integration of fault tolerance mechanisms into HLA federates and federations. The administration and execution of fault tolerant federations is supported by the so-called fault tolerant resource sharing system based on a ressource sharing system previously developed by the authors.

Bias in parallel and distributed simulation systems

Even after several decades of research, modeling is considered an art, with a high liability to produce incorrect abstractions of real world systems. Therefore, validation and verification of simulation models is considered an indispensable method to establish the credibility of developed models. In the process of parallelizing or distributing a given credible simulation model, a bias is introduced, possibly leading to serious errors in simulation results. Depending on the mechanisms used for parallelization or distribution, a separate validation of the parallel or distributed model is required. A necessary first step for such a validation is an understanding of the sources of bias that might occur through parallelization or distribution of a simulation model. The intention of this paper is to give an overview of the various types of bias and to give a formal definition of the bias and its quantification.