Bridging the Computation-Communication Gap in Future Connectivity Systems
Duke ECE faculty member Tingjun Chen adds wireless networks and systems expertise to new multi-university research center
With the number of increasingly smart, connected devices at hand, manufacturers and service providers are facing a big problem: there’s simply too much data to be efficiently processed using current infrastructure networks and information and communication technologies (ICT). What’s more, the nonstop flow of information, from devices to the cloud and back again, consumes massive amounts of energy.
The Semiconductor Research Corporation (SRC), the Defense Advanced Research Projects Agency (DARPA), and commercial partners in the semiconductor and defense sectors are co-sponsoring an initiative to develop next-generation computer architectures, called the Joint University Microelectronics Program 2.0 (JUMP 2.0).The initiative recently funded a $35 million multi-university research center that will create vast performance gains and reductions in energy consumption across massive wireless and data center systems.The new Center for Ubiquitous Connectivity (CUbiC) is led by Keren Bergman, Professor of Electrical Engineering at Columbia University.
Duke Assistant Professor of Electrical and Computer Engineering Tingjun Chen will lend his expertise in hardware, software, and algorithm co-design and wireless networking to CUbiC, as it aims to “flatten the computation-communication gap, delivering seamless edge-to-Cloud connectivity with transformational reductions in the global system energy consumption.”
“We’re really talking about developing more robust connectivity,” explained Chen. “When users move around at high speeds or in cities where buildings might block the signals of their devices, their wireless connection should still maintain fidelity and quality. The user should not notice that any switching is taking place.” That scenario may become reality through the research that CUbiCis proposing, said Chen.
Energy savings will be achieved across all the layers of the stack: more efficient devices, antennas, and transceivers, smarter algorithms for efficient resource allocation, and better system integration and optimization.
“The Center for Ubiquitous Connectivity will bring photonics onto chips, work across systems, and create new architectures, circuits, and devices that vastly expand wireless capacity,” said Bergman. “This is a new communication paradigm. Future systems will be much more energy-efficient, adaptable, secure and robust, while delivering orders of magnitude more bandwidth than we currently have.”In addition to Columbia University, CUbiC partners include the University of California, Berkeley, other participants include the University of California, Santa Barbara; University of Illinois at Urbana-Champaign; Massachusetts Institute of Technology; University of Michigan; Cornell University; Duke University; Princeton University; Stanford University; Oregon State University; University of California, San Diego; and University of Southern California.
In addition to his participation in CUbiC, Chen is also part of the Duke-led, NSF-funded AI Institute for EdgeComputing Leveraging Next Generation Networks (Athena), which focuses on edge computing—a distributed computing framework that brings applications closer to users’ devices and nearby servers.