Simulation in Emergent Computational Systems

Abstract

With the fast development of Internet and other new technologies, new computational infrastructures and paradigms are emerging. Emergent Computational Systems comprise Grid and P2P systems and wireless ad hoc network among others. The development and analysis of such systems is a challenging task due to many intrinsically complex features of these systems. Simulation appears again to be the indispensable means for modeling, performance analysis as well as implementations and analysis of algorithms before deployment in such systems.

This special issue follows the First International Workshop on Simulation and Modeling in Emergent Computational Systems (SMECS-2008) September 8-12, 2008, Portland, Oregon, USA held in conjunction with the 2008 International Conference On Parallel Processing (ICPP-08).

The papers of the workshop presented innovative methods and techniques related to Simulation, Modeling and Performance Evaluation in Emergent Computational Systems. The special issue comprises 9 papers, organized as follows.

The first paper, Recent Advances on the Simulation Models for Ad Hoc Networks: Real Traffic and Mobility Models by Doci et al., surveys recent advances on simulation models for wireless ad hoc networks. The paper stresses the importance of the mobility metrics in the simulation of wireless ad hoc networks.

In the second paper, A Simulation Framework for Dependable Distributed Systems, Dobre et al. present a solution to evaluating the correctness and performance of various dependability-related technologies for distributed systems using the formalism provided by the modeling and simulation domain. The proposed approach is based on an extension proposal of the simulation model called MONARC, which is a generic simulation framework designed for modeling large scale distributed systems.

The third paper by Doci et al., Impact of Mobility in the Connectivity of Wireless Ad Hoc Networks, presents an approach for assessing the impact of mobility in the dynamic connectivity graph, on both nodes and edges. The authors introduce an algorithm that computes the Maximum Node Degree, Link Durations, and Path Durations mobility metrics for the dynamic topology connectivity graphs. Lower and upper bounds for these mobility metrics are also provided.

In the fourth paper by Pllana et al., Automatic Performance Model Transformation From A Human-Intuitive to a Machine-Efficient Form, the authors address the issue of the development of performance models for programs that may be executed on large-scale computing systems. They propose a model to bridge the gap between the performance modeling and software engineering by incorporating UML. The user specifies graphically the performance model using UML. Then, the transformation of the performance model from the human-usable UML representation to the machine-efficient C++ representation is done automatically. The authors demonstrate the usefulness of their approach by modeling and simulating a real-world material science program.

The fifth paper by Becker et al., Replay-based synchronization of time stamps in event traces of massively parallel applications, studies the usefulness of event traces in understanding the performance behavior of message-passing applications. This study is built on earlier work by the same authors. Here, the parallel design and its implementation within the SCALASCA trace-analysis framework are presented.

In the sixth paper by Wu et al., Performance Analysis and Optimization of Parallel Scientific Applications on CMP Cluster Systems the authors present performance analysis in using CMP (Chip multiprocessors) based cluster systems for large-scale scientific applications. A detailed performance analysis to identify how applications can be modified to efficiently utilize all processors per node on CMP clusters is presented.

In the seventh paper, Baert et al., Minimization of Variance Download Times in A Distributed VOD System, examine the problem of minimizing the variance of the download time in a particular Video on Demand System based on Grid Delivery Network—an hybrid architecture based on P2P and Grid infrastructure. Different heuristics to solve it in practice, and validation through simulation are provided.

In the eighth paper by Chai et al., Virtual Large-Scale Disk Based on PC Grid, the authors develop the VLSD (Virtual Large-Scale Disk) toolkit to assist in the construction of large-scale storage using only cheap commodity hardware and software. To simulate a typical storage system in an educational environment, which may contain many small files, the authors built GCC on various storage systems and evaluated the performance. The authors show that the storage system is usable even with many small files.

In the last paper by Yang et al., Performance Evaluation of a Wireless Sensor Network for Mobile and Stationary Event Cases Considering Routing Efficiency and Goodput Metrics, the authors investigate how the sensor network performs in the case when the event node moves. They carried out the simulations for lattice topology and TwoRayGround radio model considering AODV and DSR protocols. For the performance evaluation, they considered two metrics: routing efficiency and goodput and they compare the simulation results for two cases: when the event node is mobile and stationary. The simulation results have shown that the routing efficiency for the case of mobile event node is better than the stationary event node using AODV protocol. Also, the goodput for the mobile event node case does not change too much compared with the stationary event case using AODV, but the goodput is not good when the number of nodes is increased.

We are grateful to all authors for submitting their papers to this special issue and to the reviewers for their feedback to the authors. We would like to thank Marcin Paprzycki (SCPE editor-in-chief) and Alexander Denisjuk (SCPE technical and managing editor) for the opportunity to edit this special issue and for their timely support.

Fatos Xhafa,
Department of Languages and Informatics Systems,
C/Jordi Girona 1-3, 08034 Barcelona, Spain
fatos@lsi.upc.edu

Leonard Barolli
Fukuoka Institute of Technology (FIT),
3-30-1 Wajiro-Higashi,
Higashi-Ku, Fukuoka 811-0295, Japan
barolli@fit.ac.jp