Tuesday, 12 December 2017

Tutorials

Analyzing Measurements from Data with Underlying Dependences and Heavy-Tailed Distributions

Authors:

Natalia M. Markovich (Russian Academy of Sciences)
Udo R. Krieger (Otto Friedrich University)

Abstract:

We consider measurements that are arising from a next generation network and present advanced mathematical techniques to cope with the analysis and modeling of the gathered data. These statistical techniques are required to study important performance indices of new real-time services in a multimedia Internet such as the demanded bandwidth or delay-loss profiles of packet flows during a session. The latter data sets incorporate strongly correlated or long-range dependent time series and heavy-tailed marginal distributions determining the underlying random variables of the data features. To illustrate the proposed statistical analysis concept, we use traces arising from the popular peer-to-peer video streaming application SopCast.

DOI: 10.1145/1958746.1958811

Full text: PDF

[#][]

Performance Engineering with Product-Form Models: Efficient Solutions and Applications

Authors:

Simonetta Balsamo (Università Ca' Foscari Venezia)
Andrea Marin (Università Ca' Foscari Venezia)

Abstract:

Performance engineering plays a pivotal role in the successful design of software system and the software development process. Stochastic modelling has been widely applied to predict and evaluate or estimate system performance. We consider the specification of models in terms of compositions of simpler components and their efficient solution. Various formalisms or classes of stochastic models have been applied for system performance engineering and evaluation. These formalisms includes queueing networks, Stochastic Petri Nets, and Stochastic Process Algebras. Their dynamic behaviour can be usually represented by an underlying stochastic (Markov) process. For each formalism some classes of product-form models have been identified, starting from the first remarkable results for BCMP queueing networks. For some product-form models various efficient algorithms have been defined. We discuss the problem of identifying and characterize classes of product-form models. We compare the properties of the various modeling formalisms, their solution and the combination of productform (sub)models into a heterogeneous model. We illustrate the application of product-form stochastic models for system performance engineering with some examples of tools for the solution of heterogeneous models formed by synchronized sub-models, and some practical applications.

DOI: 10.1145/1958746.1958812

Full text: PDF

[#][]

Quantitative System Evaluation with Java Modeling Tools

Authors:

Giuliano Casale (Imperial College London)
Giuseppe Serazzi (Politecnico di Milano)

Abstract:

Java Modelling Tools (JMT) is a suite of open source applications for performance evaluation and workload characterization of computer and communication systems based on queueing networks. JMT includes tools for workload characterization (JWAT), solution of queueing networks with analytical algorithms (JMVA), simulation of general-purpose queueing models (JSIM), bottleneck identification (JABA), and teaching support for Markov chain models underlying queueing systems (JMCH). This tutorial summarizes the main features of the tools that compose the suite. Furthermore, using a composite case study, we provide intuition on the versatility of JMT in dealing with the different aspects of quality-of-service (QoS) evaluation, what-if analysis, and software performance tuning.

DOI: 10.1145/1958746.1958813

Full text: PDF

[#][]