Industry Spotlight Talks

Wei Bai

Senior Researcher in the Networking Research Group at Microsoft Research Redmond

Talk Title:

RDMA at Hyperscale: Experience and Future Directions

Abstract: Remote Direct Memory Access (RDMA) has long been recognized as a powerful technology for high-performance computing and data-intensive applications. In this talk, I will present our experience in deploying intra-region RDMA to support storage workloads in Azure. Today, around 70% of traffic in Azure is RDMA and intra-region RDMA is supported in all Azure public regions. I will also discuss some open problems we see going forward and our early efforts in these directions.

Speaker Bio: Wei Bai is a senior researcher in the Networking Research Group at Microsoft Research Redmond. He received his PhD degree in computer science from Hong Kong University of Science and Technology. Wei is broadly interested in computer networking with a special focus on data center networking. His research work has been published in many top conferences and journals, such as SIGCOMM, NSDI, and IEEE/ACM Transactions on Networking. Wei also has rich experience in developing and operating production cloud networks. Currently, he is mainly focusing on high performance networking for storage, general compute and AI supercomputer.

BingYang Liu

Director of Network Technology Lab, Huawei Central Research Institute

Talk Title:

Systematic approaches to optimizing high-performance networks

Abstract: Today the demand of high-performance network is growing much faster than Moore’s Law. In this talk, I will present a series of approaches spanning from layer 7 to layer 1 that systematically improve network performance.

Speaker Bio: Bingyang Liu is the director of Network Technology Lab at Huawei Central Research Institute. His research interests now lie in data center networks and wide-area networks. Before joining Huawei, he was a postdoctoral researcher at Institute for Network Sciences and Cyberspace, Tsinghua University. He received Ph.D. from Department of Computer Science, and B.S. from School of Software, Tsinghua University.

Congcong Miao

Senior researcher at Tencent

Talk Title:

FlexWAN: Software Hardware Co-design for Flexible and Cost-Effective Optical Backbones

Abstract: The demand for WAN capacity driven by rapidly growing inter-data center traffic put forward new challenges to expensive optical networks. Today's cloud providers develop fixed optical backbone which is costly to provide WAN capacity and suffers poor performance to address failures. In this talk, I will present FlexWAN, a novel flexible WAN infrastructure to provision cost-effective WAN capacity that is resilient to optical failures. FlexWAN address system-level challenge by introducing spacing variable hardware accompanied by a centralized controller coordinating network-wide optical devices in a vendor-agnostic manner. Then, FlexWAN solves algorithmic problems with the goal of minimizing hardware costs. We evaluate the system performance in production and share insights from years of production experience.

Speaker Bio: Congcong Miao is currently a senior researcher at Tencent. Before joining Tencent, he received Ph.D. in Computer Science and Technology from Tsinghua University. He works on both underlay and overlay network and his research interests are in Optical Network, Wide Area Network, Network Verification, etc. He has published several papers on top CS conferences, including SIGCOMM, NSDI, MOBICOM, CHI.

Ennan Zhai

Director of Network Research at Alibaba Cloud

Talk Title:

Towards AI-oriented high computing power data center network

Abstract: Cloud computing scales the computing power through the data center network to meet the demand of its services for computing. However, with the rapid development of computing-intensive technologies such as AI and deep learning, the computing power of cloud has encountered unprecedented challenges. As the network performance is the performance bottleneck of computing power, it is urgent to design and develop new technologies to accelerate the network. In this talk, I will present the host-network co-design, AI-oriented data center network proposed by Alibaba Cloud as well as the corresponding technical challenges and how we design novel solution to address these challenges. These solutions mainly include congestion control approach, path scheduling approach and hardware programming approach that have been published in SIGCOMM and NSDI recent years.

Speaker Bio: Ennan Zhai is a Director of Network Research at Alibaba Cloud now. His research focuses on building high-performance and reliable networking systems. He has published more than 30 papers in top-tier conferences such as SIGCOMM, NSDI, and OSDI. He won the first prize of Technology Invention Award of China Institute of Communications. Prior to joining Alibaba in Jun 2018, he was a research scientist and lecturer in the Computer Science Department at Yale University where he received his Ph.D. degree in 2015.