Invited Talks

• A Dynamically Configurable Components Platform
  Bertil Folliot (University Paris 6)

• Naming Schemes for Wireless Ad-hoc Clusters
  Dan Grigoras (UCC Cork, Ireland)

  abstract (PDF) · presentation (PDF)

• The challenges and the promise of quantum computing and quantum information theory. Reflections on quantum parallelism
  Dan C. Marinescu and Gabriela M. Marinescu (University of Central Florida, Computer Science Department)

  abstract (PDF) · presentation (PDF)

• A Proposal and Evaluation of the Fast Reconnect Ad-Hoc Network Routing Protocol
  Kazuhiro Mizoguchi et al. (Kyushu University, Japan)

  presentation (PDF)

• Communicating Systems of Diffusing Mobile Objects
  Traian Muntean (CNRS-Grenoble & University Méditerranée-Marseille)

  This talk will sketch out our work in the field of diffusing communicating systems mainly targeting correct construction of such systems. I shall insist, in particular, on:

  • a new process algebra for broadcasting objects and mobility

  • adaptive routing for mobile objects

  • a distributed refinement methodology for automatic construction of such systems

  presentation (PDF)

• Distributed Multimedia Streaming in a Heterogeneous Environment
  Radu Cornea, Shivajit Mohapatra, Nikil Dutt, Alex Nicolau, Nalini Venkatasubramanian (Dept. of Information & Computer Science, UC Irvine)

  presentation (PDF)

• Resource Bartering in Grids
  Can Ozturan (Dept. of Computer Engineering -         Bogazici University, Istanbul, Turkey)

  Grids will allow various resources to be shared among many users. This sharing however will definitely not mean that everyone will have unrestricted use of the resources. Some mechanism such as pricing or quotas can be employed in order to enforce controlled sharing of resources. In this work, we propose a barter model for resource sharing whereby users or computer centers can get something in return for letting their resources to be used by others. We develop a barter model in which multiple resources can be traded by making use of directed hypergraphs.  We prove that the decision version of the multi-item bartering problem is NP-complete. We present an integer programming formulation for the bartering problem. We also present a linear time algorithm to compute components that may contain feasible bartering solutions. Finally, we present various computational results from our software that makes use of LP_SOLVE and CPLEX mixed integer programming libraries to solve example bartering problems.

  presentation (PDF)

• Transaction Processing for Clustered Virtual Environments
  Christian Shallhart (Computer Science Department, Vienna University of Technology)

  presentation (PDF)

• Modeling Resource Management in Concurrent Computing Systems
  Isaac Sherson (UC Irvine)

  presentation (PDF)

• A Compositional Semantics for Petri Nets
  Ferucio Laurentiu Tiplea and Aurora Tiplea (Faculty of Computer Science, "A.I.Cuza" University of Iasi, Romania)

  abstract (PDF)

• Dynamic SMP Clusters as the Architectural Solution for Fine-Grain Numerical Computations
  Marek Tudruj (Polish-Japanese Institute of Information Technology, Warsaw, Poland)

  abstract (PDF)

• Optimizing Object Oriented Programs Based on the Bytecode-Defined Data Dependence Graphs
  Marek Tudruj, Erik Laskowski, Bernard Toursel, Violeta Felea

  The talk is concerned with the scheduling of the bytecode generated for parallel object oriented programs written in Java. Two optimisation methods are proposed based on the use of several levels of graph program representation. Class dependence graphs are used to depict class inheritance in a program. Method dependence graph (call graph) is used to show intensity of inter-method calls. The first proposed optimization heuristics uses method dependence graphs to preserve locality of intra-threads method calls and to reduce number of remote object method calls. Based on the bytecode generated for the program, an extended macro dataflow graph can be generated which depicts data dependencies between macro data nodes in threads of the program. The second optimization algorithm applies the macro data node clustering algorithm to the graph. It defines distribution of nodes and objects on processors that tries to minimize the total execution time of the program. Object clustering method is based on the DSC approach.

  presentation (PDF)