IEEE International Conference on High Performance Switching and Routing
26-29 May 2019 // Xi'an, China

Keynotes

KEYNOTE 1 – Monday May 28, 09:40 – 10:25

Chongjin Xie, Ph.D., Senior Director, Chief Communication Scientist, Network Research Division, Alibaba Infrastructure Service, Alibaba Group

Dr. Chongjin Xie is a senior director and chief communication scientist in AIS, Alibaba Group, leading an optical network team to develop datacenter optical technologies to support Alibaba online platform and cloud services. Prior to joining Alibaba in 2014, Dr. Xie was a Distinguished Member of Technical Staff at Bell Labs, Alcatel-Lucent, doing research on optical communication systems and networks. He did his postdoctoral research at Chalmers University of Technology in Sweden from 1999 to 2001, and received his M.Sc. and Ph.D. degrees from Beijing University of Posts & Telecommunications in 1996 and 1999, respectively. Dr. Xie has published one book, 4 book chapters and over 200 journal and conference papers. He was a program chair of OFC’2019, and has been serving as chairs, TPC chairs or TPC members in many conferences and an associate editor of Journal of Lightwave Technology. Dr. Xie is a Fellow of OSA and a senior member of IEEE.

Title: Optical Technologies for Hyper-Scale Cloud Datacenters

Hyper-scale cloud datacenters, where most of the Internet services such as e-commerce, video streaming and social networking are running, have become the fundamental infrastructure for cloud computing and big data services. The fast-growing demand for computing and storage power, which is provided by massive numbers of powerful servers connected together in data centers, requires scalable and efficient datacenter networks. Most of the interconnects in today’s hyper-scale datacenter networks, including both short-reach and long-reach interconnects, use optical technologies. The demand for large volume, fast traffic growth and short technology life cycle optical interconnect technologies present a big challenge for optical communication industry. In this talk, we will present requirements and challenges of optical technologies for hyper-scale datacenters, including both short-reach optical interconnects and long-reach optical transport technologies. Potential solutions to solve these challenges are discussed.

 

KEYNOTE 2 – Tuesday May 28, 08:40 – 09:25

David Neilson, Ph.D., Technical Manager, Nokia Bell-Labs

Dr David T. Neilson is Technical Manager and Bell Labs Fellow, Hybrid Integrated Devices Research Department, Bell Labs, Nokia, Holmdel NJ USA. He joined Bell-Labs in 1998 where he leads research efforts on devices systems and architectures for optical switching and optical interconnects for electronic switching. Over the last 20 years at Bell Labs he has researched several optical switching systems and technologies including using MEMS and LCoS for wavelength selective switches and optical crossconnect. He has also led groups working on photonic integration in III-V and silicon materials.  Prior to joining Bell Labs, was a Visiting Scientist at NEC Research, Princeton NJ, researching optical interconnects for high performance computing systems. He has authored over 200 publications and patents on both devices and systems in the field of optical interconnects and switching. He has chaired and been on several IEEE-LEOS, OSA, and SPIE conference programs committees, in the field of optical interconnects and switching.  He is a Fellow of IEEE.

Title: Future switching requirements for long haul and metro networks

Optical networks serve as the cornerstone of our connected society, enabling the information superhighway that delivers the Internet all across the globe.  As the number of users and data services increase, the network technology and architecture must adapt for it to continue to efficiently and economically support the larger traffic loads. Optical communication systems, have demonstrated equivalent per fiber capacity scaling over the last decades to match traffic load and today’s commercial long-haul systems can carry 50Tb/s on a single fiber. However, we have now reached the limit of the capacity on a single fiber and it is anticipated that to continue to achieve the anticipated 40% pa growth rate in capacity it will require the use of many more parallel fibers. Driven by this change in fiber infrastructure and the larger transported capacity, the network switching elements and their functionality must also evolve.

 

KEYNOTE 3 – Wednesday May 29, 08:40 – 09:25

Prof. Chengchen Hu, Principal Engineer, Director of Xilinx Labs Asia Pacific

Chengchen Hu is a Principal Engineer at Xilinx Inc., world leader and the inventor of the FPGA (Field-Programmable Gate Array) and the ACAP (Adaptive Compute Acceleration Platform). He is the founding director of Xilinx Labs Asia Pacific based in Singapore currently leading research on networked processing systems. Prior to joining Xilinx in Aug., 2017, he was a Professor and the Department Head at the Department of Computer Science and Technology, Xi’an Jiaotong University in P. R. China. He received Ph. D. degree in Computer Science from Tsinghua University and is recipient of the New Century Excellent Talents in University award from Ministry of Education, China, a fellowship from the European Research Consortium for Informatics and Mathematics (ERCIM), a fellowship of Microsoft “Star-Track” Young Faculty. He severed in the organization committee and technical program committee of many conferences including INFOCOM, IWQoS, ICC, GLOBECOM, ANCS, Networking, APCC, etc.

Title: Network Hardware and Software Co-Design for Data Centers

Cloud computing, which is very successful both technically and commercially, has been evolving from enterprise data centers to warehouse-scale data centers.  Distributed computing applications are the major driver of the cloud evolution.  Distributed systems have existed in a loosely-coupled, CPU-centric style for several decades, but their performance improvements are now declining, due to the slowing of Moore’s law, explosive data growth, and unprecedented computational demands.  Increasingly, near-storage compute and in-network compute are becoming the way forward, leveraging specialized network hardware and software to provide reliable, predictable, efficient and high-performance infrastructure.  The talk will cover recent Xilinx research and development on data center hardware and software co-design from a networking perspective. And it will also discuss further directions that should be investigated next.