Content Caching and Delivery: Fundamental Limits, Challenges, and Opportunities
Designing the infrastructure to handle massive amounts of data requires tools and techniques from different research areas including communication, computing, and storage. Each of these areas has different performance metrics such as rate, delay, storage and bandwidth costs. Traditionally, a subset of these metrics is studied in isolation within each research area. In this talk, we argue that these metrics should instead be analyzed jointly by characterizing the fundamental trade-offs between them.
In the first part of the talk, we focus on the fundamental trade-off between communication and storage in cache networks. The today's caching algorithms optimize the caches in order to maximize the amount of content delivered locally. We show that this approach is inefficient for cache networks. We introduce a new formulation of the caching problem, focusing on its basic structure. For this setting, we propose a novel coded caching approach that achieves significantly larger reduction in peak rate compared to previously known caching schemes. In particular, the improvement can be on the order of the number of caches in the network. This result leads to the first fundamental characterization of the rate-memory trade-off for systems with more than one cache.
In the second part of the talk, we shift focus to the interplay of network delay and communication rate. As a representative example, we consider the emerging application of offloading of base-band processing from wireless networks to the cloud. Network delay could prevent the use of traditional approaches of interference management such as cooperative interference cancellation and interference alignment. In this talk, we propose a new approach for interference management that achieves a rate close to cooperative schemes, but that is tolerant of even large delays.
Mohammad Ali Maddah-Ali received the B.Sc. degree from Isfahan University of Technology, the M.A.Sc. degree from the University of Tehran, and the Ph.D. degree from University of Waterloo, all in electrical engineering. From 2007 to 2008 he worked at the Wireless Technology Laboratories, Nortel Networks. From 2008 to 2010, he was a post-doctoral fellow at University of California at Berkeley. Since 2010, he has been working as a research scientist at Bell Labs, Alcatel-Lucent, Holmdel, NJ. Dr. Maddah-Ali received several awards including the NSERC Postdoctoral Fellowship in 2007, official mention from the IEEE Information Theory Society for introducing interference alignment in 2009, and the best paper award at IEEE International Conference on Communications (ICC) in 2014.
Prof. Salman Avestimehr