Addressing the Deluge of Data and Connectivity Bottlenecks

The Center for Ubiquitous Connectivity recently held its inaugural annual review at Columbia to discuss research and collaborations advancing energy-efficient communications technologies.

In a world where every swipe on the phone, every mobile payment, and every online interaction involves data exchange, connectivity emerges as a necessity. The idea behind “ubiquitous connectivity,” a state where devices effortlessly create, share, and process data, and never lose connection, is increasingly–and seamlessly–a  part of our everyday lives. 

At the inaugural annual review of Columbia Engineering’s Center for Ubiquitous Connectivity (CUbiC), held June 27-28, faculty, students, researchers, industry experts, and government officials convened to discuss new and ongoing research and share insights, fostering collaboration and driving advancements in the field of ubiquitous connectivity. Attended by about 115 in-person guests and additional participants who joined virtually, the two-day event at Columbia’s Morningside campus delved into three research themes: connectivity networks and systems, wireline and lightwave interconnects, and wireless circuits and technology. 

CUbiC, a research center led by Columbia Engineering in collaboration with 12 other academic institutions, launched in January with $35 million in grant funding by the Joint University Microelectronics Program 2.0 (JUMP 2.0). The Center is part of an initiative to support high-risk, high-payoff research that addresses existing and emerging challenges in information and communication technologies, co-sponsored by the Semiconductor Research Corporation (SRC), the Defense Advanced Research Projects Agency (DARPA), the commercial semiconductor industry, and the defense industrial base. 

In an explosive data age, deluged data centers have met significant challenges to keep up with the data hunger of wireless devices. The goal of CUbiC is to advance energy-efficient communications technologies to address these bottlenecks. Over the next five years, CUbiC will strive to flatten the computation-communication gap, delivering seamless “edge-to-cloud” connectivity with transformational reductions in the global system energy consumption. The "edge" refers to devices or systems that gather data, such as sensors or local computing devices, while the “cloud" are the remote data servers that offer storage, computing power, and services over the internet. The concept of edge-to-cloud involves the flow of data and services between these two components. 

Reflecting upon the rich legacy of Columbia’s expertise in the fields of telecommunications, wireless communications, and photonics, Columbia Engineering Dean Shih-Fu Chang said in welcome remarks that CUbiC, and other important collaborative research partnerships that the School has recently established, is key to the School’s success in developing and scaling new technologies.

In an overview of the center, CUbiC Director Keren Bergman said, “Our goal here is to scale performance, while reducing energy by at least orders of magnitude.” Bergman is spearheading research in the development of nanoscale photonic interconnect technologies. These technologies aim to enable seamless data transfer across on-chip networks, memory systems, and large-scale computing systems with exceptional energy efficiency. 

Bergman also highlighted CUbiC’s diverse team of experienced senior researchers, early career researchers, and an extensive student body committed to broadening participation and improving diversity in the field overall. CUbiC now has nearly 89 SRC graduate students and 10 postdoctoral fellows, with 20% constituting underrepresented minorities in the field.

Within each theme, the conference featured three engaging presentations from industry partners and university researchers, offering diverse perspectives on the latest developments. Following the talks, a poster session was held as a platform for students to showcase their innovative projects and engage in fruitful discussions with industry partners and researchers about their work.

Young investigators spotlight

CUbiC student researcher, Lalitha Giridhar, who is a doctoral student in electrical and computer science at University of California at Santa Barbara, is focusing on millimeter wave (mmWave) line-of-sight (LoS) multiple-input multiple-output (MIMO) communication systems. At the poster session, she presented “Creating Spatial Degrees of Freedom for Long-Range LoS MIMO using Reflect-arrays.”

Reflect-arrays are typically planar arrays consisting of low-cost reflecting elements which can be controlled individually to cooperatively achieve a specific purpose. As nearly passive elements, reflect-arrays can be deployed ubiquitously in a LoS MIMO communication link. LoS MIMO technology enables achieving high data rates comparable to optical links. Giridhar’s project aligns with the demands of the next generation of technology, including 5G cellular networks. By leveraging the advantages of LoS MIMO and reflect-arrays, it becomes possible to boost the current performance and enable higher data rates while retaining compact transceiver form factors, according to Giridhar. Her findings emphasize the potential of this technology for addressing the needs of the industry while ensuring efficient resource allocation.

Giridhar said that having the opportunity to collaborate with industry experts at CUbiC is a huge plus. “That is really what CUbiC is about: constructive collaboration between academia and industry,” said Giridhar. CUbiC provides the perfect environment, she said, for young researchers to interact with industry to effectively move forward on new technology.

Ji Yoon Han, another CUbic student researcher who participated at the poster session, is an incoming PhD student at University of Michigan who completed her master's degree in electrical and computer engineering in May. Han’s primary focus of research is in designing systems-on-chip (SoC) for advanced communication systems. Her work combines theory, algorithm development, and hardware implementation to develop innovative solutions in this field. 

Han presented her research, “Beam Space Processing for mmWave Massive MU-MIMO Communication Systems,” and notes that “it is an honor” to be a student researcher at  CUbiC. “CUbiC gives us a great opportunity to connect with the top researchers and industry [leaders] in the field, and have their thoughts on our work,” she said.

The goal of Han’s  project is to develop a DSP for a massive multi-user MIMO system, which adapts to the channel condition, antenna configurations and the front end. To mitigate the complexity, one of the critical challenges in the mmWave multi-user MIMO system, she proposes a beam space-based MIMO system that leverages the sparsity of the channel. By converting to the beam space domain processing from the conventional antenna space domain processing, it is possible to reduce complexity without minimizing the performance. Her preliminary results demonstrate the feasibility of this approach. “I think this definitely helps to realize the low latency and reliable wireless connectivity with massive capacity, which is one of the center’s goals,” she said. 

Following the formal presentations and poster session, participants were invited on an informative tour of the laboratories, gaining insights into the cutting-edge research and technology advancements happening at Columbia Engineering. Students had the opportunity to engage in meaningful networking sessions, connecting with industry professionals and peers to expand their professional network. Additionally, the industry caucus session and CUbiC industry panel discussion featured experts who shared valuable insights and discussed key trends in the field.

Read original article