It is a pleasure for me to welcome you at this special issue of the International Journal of Parallel, Emergent and Distributed Systems, which collects some of the best papers presented at the Workshop on Dependable Parallel and Network – Centric Systems (DPDNS08), held in Miami in April 2008.
The workshop program focused on all aspects of network computing, dependability, wireless communications, fault tolerance, covering comprehensive topics in research and applications, addressing and discussing all the issues related to the converging fields of dependable computers and communications.
This special issue includes four papers that address performance and reliability modeling issues of complex dynamic systems, covering both theoretical aspects and their application to concrete case studies. The first two papers address the problem of modeling complex dynamic systems even through approximation techniques, while the remaining two more specifically focus on case studies belonging to the fields of wireless networks and software development.
Romano et al. show how the measures of interest for the MMPP/M/1 queue can be approximated by a weighted sums of corresponding M/M/1 queues, with state dependent arrival rates adopted from the MMPP. By jointly considering both accuracy and cost, they identify the scenarios where such approximate solution can be effectively used in support of network servers (dynamic) configuration and evaluation strategies aimed at ensuring agreed dependability levels in case of, e.g., request redirection due to faults. A real-world case study is finally evaluated, relying on a trace based traffic characterization of a Grid server.
Distefano describes dynamic reliability block diagrams (DRBD), a novel technique for modeling the dependability of fault-tolerant systems. In order to demonstrate the effectiveness of the proposed technique, he mainly focuses on fault tolerant-dynamic-dependent computing systems, investigating some dynamic reliability/availability behaviors and providing the guidelines for their representation and evaluation. The discussion is supported by the results obtained evaluating a complex fault tolerant computing system with several units affected by common cause events, shared workloads and other dynamic-dependable behaviors.
Lucas et al. provide an integrated approach to improve dependability in wireless mesh networks (WMN). Their approach combines network coverage planning on the physical layer, bandwidth management on the link layer and live network monitoring to improve the reliability, availability and maintainability of a WMN. They provide fine grained means to improve the predictability of the network components, thus making the WMN more dependable.
Performance related problems are a key aspect in the Software Development Process (SDP). In particular an early integration of performance considerations inside the SDP has been recognized during last years as an effective approach to speed up the production of high quality software. In this paper Scarpa et al. introduce a design process that can help in producing high reliable software with predefined performance requirements. They present the performance plug-in ArgoPerformance able to help a software designer in a Software Development Process, by setting and retrieving back performance indices from a UML model described inside the ArgoUML tool.
The main conclusion I wish to raise to your attention, is that all the papers make wide use of modeling to evaluate performance and dependability aspects of complex systems. This demonstrates once more how important analytical modeling and simulation are to analyze and predict the behavior of modern systems, that by their nature are growing in complexity, due also to the increasing number of their interacting components. Being able to properly manage such techniques is extremely important and can dramatically contribute to improve the quality of the entire development process, from the design phase, to the implementation and the management ones. Software systems are becoming more complex, and involve and increasing number of system's functionalities. Software modeling is then a relatively new research area, to be further investigated due to the crucial role it plays in determining the dependability and performance offered by the whole system.
I hope you will find this special issue of interest for your immediate needs and stimulating towards new research interests. I would like to express my appreciation to the authors of all the papers, without whom this special issue would not have been possible, and to the expert reviewers who worked diligently under the usual tight schedules.
I would also like to thank Dr Ivan Stojmenovic, the Editor of IJPEDS, for his continuous and valuable support in managing all the details required in the smooth organization of this special issue.
Antonio Puliafito
University of Messina