Ogras’ networks-on-chip paper stands the test of time

// Electrical & Computer Engineering

Photo of Umit Ogras
Umit Ogras

While finishing his PhD at Carnegie Mellon University in 2005, University of Wisconsin-Madison Associate Professor of Electrical and Computer Engineering Umit Ogras was co-author on a paper about networks-on-chip (NoC) architecture. Now, 15 years later, the paper has received a Test of Time Award at the virtual Embedded Systems Week, a joint conference held by several societies and corporate sponsors.

The Test of Time Award honors academic papers that still contribute to their field years or decades after first being published. Ogras’s paper is titled “Key research problems in NoC design: a holistic perspective,” and appeared in the Proceedings of IEEE/ACM/IFIP International Conference on Hardware/Software Codesign and System Synthesis in 2005. Co-authors include Radu Marculescu and Jingcao Hu, also of Carnegie Mellon.

Networks on chips were first proposed in the early 2000s. Ogras believes the paper made an impact because at the time of its publication there was a pressing need for NoCs to handle the increasing performance demands from computing systems and multi-core chips. Work on NoCs eventually moved from academia to the industrial sector, which has continued fueling research in this area.

He also says the paper was not written as a classic survey. While it reviewed current solutions and problems in NoC architecture, it also proposed a new taxonomy of key research problems. It included a mathematical formulation of each issue, making it easier for other researchers to engage with the outstanding problems.

“One of our goals with Radu Marculescu was to contribute to the field by catalyzing NoC research through a systematic study,” Ogras says. “Looking at the impact of this paper, rich literature on NoCs and industrial adoption, I feel that we met our long-term goal. This is satisfying since measuring long-term goals is not always possible.”

Author: Jason Daley