People

Ibrahim Matta Assistant Professor PhD, University of Maryland at College Park
matta@ccs.neu.edu Personal Home Page

Professor Matta, who joined the college in 1995, conducts research on resource allocation in large, high-performance, heterogeneous distributed systems, particularly the design and evaluation of integrated-services (multimedia) networks. He focuses on the problem of routing diverse traffic streams with diverse quality-of-service (QoS) requirements.

Professor Matta is seeking to develop highly efficient, robust, and scalable QoS routing protocols. Currently, he is developing fast, near-optimal heuristics that employ innovative resource allocation techniques and are based on integrated dynamic-control models. These models consider the time-dependent interplay between routing and other QoS controls employed at different levels of the network architecture. Therefore, routing heuristics generating near-optimal unicast and multicast QoS paths very rapidly through QoS-aware path construction and search-space reduction techniques can be developed. With a National Science Founda-tion (NSF) Early Career Development grant, Professor Matta is investigating a new "load profiling" view of route assignment, in which the distribution of available resources on various routes is maintained as close as possible to the distribution of QoS demands.

Cofounder of the college's Laboratory for Networking and Distributed Computing, Professor Matta is also establishing another experimental network laboratory, recently funded by an NSF Major Research Instrumentation award.

Career Publication Highlights

Matta, Ibrahim. 1995. Transient and steady-state performance of routing protocols: Distance-vector versus link-state. Journal of Internetworking Research & Experience 6:59-87.

---. 1995. Type-of-service routing in datagram delivery systems. IEEE Journal on Selected Areas in Communications: Special Issue on the Internet 13, no. 3:1411-1425.

Matta, Ibrahim, and A. Udaya Shankar. 1995. Z-iteration: A simple method for throughput estimation in time-dependent multiclass systems. In proceedings, ACM Sig-metrics/Performance '95, 126-135. New York: ACM Press.

Research Home Page