Sotirios G. Ziavras - US grants

The main objective of this research is to solve the problem of mapping application algorithms developed for multilevel architectures, which are more general than the standard-pyramid architecture, onto hypercube-based systems. Hypercube-based systems are commercially available and have widely been used by the scientific community because they are highly flexible, fault-tolerant, and efficient. Researchers have been trying with success, to solve the problem of embedding standard pyramids into hypercubes. However, such embeddings do not guarantee very high performance for all of the multilevel algorithms, because the performance requirements of particular algorithms are not taken into account by the embedding process. This project will evolve through five phases. The first phase will propose algorithms for embedding multilevel structures into hypercubes. The second phase will be involved in the refinement of the embedding algorithms proposed in the first phase, in such a way that the exact computation and communication requirements of the application algorithms should affect the embeddings. The third phase will attempt modifications of the embedding algorithms proposed in the second phase, so that they will be taking into account statistical performance data, in order to become suitable for the mapping of non-deterministic and data-dependent multilevel algorithms. The fourth phase of this project will propose a class of "multicube" systems, which, while providing smaller numbers of communication channels, can yield performance comparable to that of the hypercube. To conclude techniques for mapping multilevel algorithms onto hypercube-based and multicube systems will be developed. Analytical techniques and simulation results will be used to prove the effectiveness and efficiency of the mapping algorithms.

9500260 Manikopoulos The main objective of this project is to complete the existing Turbonet multi-processor prototype system; this will enable the researchers to investigate how to combine effectively the shared memory and message passing paradigms in a realistic multi-computer structure for selected classes of algorithms. Turbonet is a prototype system which physically implements both paradigms. It is a desirable system, in terms of its cost/performance ratio, to utilize in many computationally challenging applications of practical significance. This work is aimed to alleviate the weakness of parallel systems based solely on shared memory or message passing. An additional important issue to be aided in this work is load balancing which is a major problem in a growing number of scientific applications that give rise to programs with dynamic execution profiles.

In addition to the traffic growth in the Internet, new applications and services are imposing a large variety of requirements. At the same time, the emergence of network applications on video, audio, business services, and other data ser-vices is creating an imminent demand for integration of most traffic with de-fined quality of service (QoS) guarantees. Examples of emerging network re-quirements are directly related to reliability, recoverability, and security. Al-though most of these requirements have kindled research interests in recent years, it is still unclear how to integrate them for next generation networks. To manage the increasing variety of QoS requirements, this research project pro-poses a new service model concept, called service vector, as a solution for pro-viding QoS support for a large variety of traffic classes. This concept is expected to help achieving the following objectives: a) robust differentiated service model capable of supporting fine QoS granularity; b) scalability; c) satisfaction of the users' customized end-to-end requirements; d) improved network operator revenue; e) higher utilization of the current Diffserv network model. The com-plexity and feasibility of service vectors for the next generation networks is in-vestigated, and the impact on router architectures and deployment issues are considered for the implementation of service vectors in new-generation and de-ployed routers.

The National Science Foundation (NSF) Graduate Research Fellowship Program (GRFP) is a highly competitive, federal fellowship program. GRFP helps ensure the vitality and diversity of the scientific and engineering workforce of the United States. The program recognizes and supports outstanding graduate students who are pursuing research-based master's and doctoral degrees in science, technology, engineering, and mathematics (STEM) and in STEM education. The GRFP provides three years of financial support for the graduate education of individuals who have demonstrated their potential for significant research achievements in STEM and STEM education. This award supports the NSF Graduate Fellows pursuing graduate education at this GRFP institution.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.