Date(s) - 20/05/2014
14:30 - 15:30
Talk: We will present some of our recent results on cellular networks. More precisely, we want to show how the operation of cellular networks has to be revisited in the following two cases. The first case deals with homogeneous networks where base stations can be switched off to save energy. The second case focuses on heterogeneous networks with limited backhaul.
Base Station Switching (BSS): Base station operation consumes a significant amount of energy, a considerable amount of which can be saved by switching off base stations during low user demand (for example, at night). BSS is a simple idea but yields complex operational challenges that we address in a comprehensive manner. The first challenge is linked to energy saving, i.e., the design of BSS patterns at a global system level. We propose a set of BSS patterns and study their characteristics. The second challenge is to ensure coverage both on the uplink and the downlink. We show that coverage is mainly an uplink issue and requires careful transmit power management. The third challenge is user performance which is critically linked to scheduling and interference management. We show that existing scheduling schemes do not work when BSS is used and we then propose novel scheduling schemes for both uplink and downlink that take interference into account in a realistic fashion, ensure coverage and provide good energy-performance trade-off for the proposed BSS patterns.
Limited Backhaul: we investigate the impact of limited backhaul capacity on user scheduling in the context of a heterogeneous network comprising a macro base station overlaid with small-cells. When user association and channel allocation are given, like in the infinite-capacity backhaul case, the global proportional fair (PF) user scheduling problem can be decomposed into a set of independent local PF user scheduling problems. However, a local PF scheme with finite backhaul, unlike the case with infinite backhaul, does not always give equal time to each user. The results show that a backhaul-aware scheduling scheme is simple to implement and necessary to obtain the best performance.
This work was done in collaboration with A. Kumar and J. Ghimire.
Bio: Catherine Rosenberg is a Professor in Electrical and Computer Engineering at the University of Waterloo. Since June 2010, she holds the Tier 1 Canada Research Chair in the Future Internet. From 1999 to 2004, Prof. Rosenberg was a Professor in the School of Electrical and Computer Engineering at Purdue University. Prof. Rosenberg is on the Scientific Advisory Board of Orange (formerly France-Telecom) and of IRT<>bcom. She was elected an IEEE Fellow for contributions to resource management in wireless and satellite networks on 2011 and a Fellow of the Canadian Academy of Engineering in 2013. Her research interests are mainly in two areas: Wireless Networks and Energy Systems.