题目：Stochastic Network Calculus: A fundamental insight on the basic relationship between demand, resource and achievable performance in computer networks
报告人：Dr. LIU Yong
It has been found very difficult to apply traditional queuing theory for performance analysis and resource allocation for Internet. Most of theoretical results from queuing theory are only for single node analysis while for Internet we are more interested on end-to-end results over an entire network. In addition, most of traffic distribution models such as Poisson are not valid for Internet traffic. In viewing of this, we need a fundamental insight on the basic relationship between demand, resource and achievable performance in computer networks.
In this talk, a new approach, referred to as stochastic network calculus, is introduced to systematically investigate the relationship between traffic demand, system resource and achievable performance in computer networks. In this theory, traffic arrival process is modelled by a stochastic arrival curve. System resource (such as routers, switches) is modelled by a stochastic service curve. With these models, traffic departure curve can be obtained by convolution of the stochastic traffic arrival curve and stochastic service curve. Various achievable performance metrics such as backlog and delay can be obtained as well. More importantly, the whole network can be treated as a virtual server with convolution of all servers in tandem, such that the input-output characterization of the whole network can be obtained as well. This kind of input-output characterization is similar to signal process system or control system theory in which output function of a system can be obtained by a certain transformation of input function and system function. This theory can be considered as an information theory for computer networks as it provides fundamental insight to the basic relationship between resource and demand. Due to this reason, the theory has been widely applied to various areas such as performance analysis of wireless networks, video-on-demand over peer-to-peer networks, network coding, and even outside the area of computer networks such as energy generation in power grid, etc. It can also be used to answer some fundamental questions in the area of cloud computing such that what kind of service level can be achieved with the given system resources or how many resources are required to achieve certain level of performance.
Dr. Liu Yong holds a PhD degree from National University of Singapore. After PhD study, he has worked as Post-Doc researcher at Royal Swedish Academic of Science. After that, he has worked at National University of Singapore and Data Storage Institute under A*STAR at Singapore. He has received multiple awards and research grants from Singapore government agencies such as EDB and SPRING Singapore. His research areas include computer network performance evaluation, quality of service provisioning, resource allocation for cloud data centers, optical networks, network service protection and fast restoration. He has published numerous high quality papers on IEEE Transaction Journals. He is also the co-author of two books in computer networks “Stochastic Network Calculus”, “Quality of Service in Optical Burst Switched Networks”. The stochastic network calculus theory that Dr. Liu established has been cited and applied widely to numerous areas such as performance analysis of wireless networks, video-on-demand in peer-to-peer networks, network coding, etc. It has been also applied to areas outside computer networks, such as performance evaluation of supply and demand process of renewable energy in power grid.
He is the CTO and co-founder of a high-tech company, Jitcomm Networks, which is dedicated on R&D on high performance network optimization and management equipment. The company has received numerous funding from National University of Singapore, SPRING Singapore and Singapore-based Venture Capital firm. The company products have been sold to universities, schools, government agencies, hotels, marine shipping companies, multi-national companies at China, Singapore, Malaysia, Thailand and Europe markets. He has been awarded as China-Singapore Suzhou Industrial Park Science & Technology Leader at 2011.