Fwd: In Egypt during December
Mohammed Nafie
Dynamic Spectrum Allocation (DSA) for Cellular Access Networks”
Several studies,initiated recently by the US regulator FCC,have shown that the frequency spectrum is inefficiently exploited:
some bands are highly crowded (at some day hours or in certain dense urban areas) while others remain poorly used. This situation,together with the rapid evolution of the SDR (Software Defined Radio),has led to the development of opportunistic CR (Cognitive Radio) systems on one hand,and DSA (Dynamic Spectrum Allocation/Access) systems on the other hand.
In this talk,we focus on DSA systems and adopt the cellular operators point of view to propose and evaluate DSA algorithms. The proposed algorithms are all based on the idea of sharing a common pool of spectral resources between different operators or different access networks. The work focuses on packet traffic services while both technical and pricing aspects are taken into consideration. The idea of sharing resources has led us to explore competition and cooperation aspects between operators. These aspects are referred to the costs and revenues partitioned among the operators as a result of spectrum sharing. There is indeed a clear trade-off for an operator between the spectrum cost and the revenues obtained from the end-users. When inter-operator DSA is considered,it is also difficult to separate technical from pricing aspects, especially for operators who pay high prices for the license. In this work, we thus propose several utility functions to be maximized taking into account spectrum price and end-user QoS (quality of service).
Several scenarios are studied. We adopt centralized as well as distributed models while designing algorithms for DSA and we investigate both mono-operator and multi-operator cases. We also study two possibilities for the periodicity of the allocation process instant:the spectrum allocation process is performed periodically,or at system state change (i.e. at each arrival or departure of end-users).
Our algorithms first take advantage of temporal heterogeneities of the traffic with the aim of a more efficient usage of the spectrum. This efficiency is obtained by sharing resources between different networks and adapting the amount of allocated spectrum to the network loads. The proposed DSA algorithms are based on different methods:SMDP (Semi Markov Decision Process),Q-learning, Game Theory and heuristics. Each approach is adapted to one of the studied scenarios. The way DSA algorithms make use of the spatial heterogeneity of the traffic is our last research axis. We investigate the dynamic spectrum allocation problem for one cellular operator who leases spectrum from the regulator,in a spectrum sharing context. Here,we focus on the allocation of spectral resources to individual cells of a network. Two difficulties arise. On the one hand,interference has to be taken into account because it locally modifies the profit that can be expected from a frequency block. On the other hand,as the network may be potentially very large, distributed algorithms are needed. In this work, we propose an approach based on Tabu search to cope with theses issues.
We show that,for the different studied scenarios, the DSA algorithms proposed in this work allow the operator to gain more rewards than with the traditional FSA (Fixed Spectrum Allocation). The DSA algorithms also reduces the spectrum usage while maintaining an acceptable end-user QoS. In some cases,it is shown that they achieve a lower blocking probability. DSA gains are particularly significant at low and moderate load conditions of the networks. They are however obtained at the price of a reduced users’ throughput. At high loads,DSA and FSA achieve similar results.Best Regards,
Hany
Hany K. HASSAN received his PhD degree on Dynamic Spectrum Access (DSA) techniques for cellular systems, from the “Computing and Networks” department at Telecom ParisTech, France, 2010. He has received his B.Sc. degree in communication engineering from Alexandria University, Egypt, on 1999. He received his M.Sc. from TELECOM ParisTech, France, in 2002, where he worked on channel estimation for wireless LAN systems. After his masters, he worked on overload algorithms for 3G networks within the R&D team at Alcatel (Velizy). He worked as a radio design engineer in Alcatel (-Lucent) mainly on indoor design projects for 3G and 2G cellular systems. He also joined the WiMAX Advanced Radio Engineering (ARE) group within the company late 2006. Hany Kamal has co-authored several methodology-documents within Alcatel-Lucent. He also co-authored several deliverables within the French research project Urban Planning for Radio Communication (URC). His current research interests are in wireless networks, dynamic spectrum access techniques and in Intelligent Transportation Systems (ITS)