Search

The purpose of the Wireless Regional Area Networks (WRANs) is to allow the use of spectrum allocated for television broadcasting in areas with sparse users, ensuring no interference with licensed operators. Cognitive radios using the IEEE 802.22 standard should monitor the environment to not cause interference in primary operators. When multiple WRANs intend to operate on the same set of frequencies in overlapping areas, the 802.22 standard provides mechanisms for their co-existence and provides a temporal multiplexing scheme to solve this problem. We propose three channel allocation algorithms to support the co-existence of potentially interfering WRANs. The proposed algorithms are based on graph coloring, and the innovation is to consider the demand for bandwidth from each WRAN, maintaining proportionality between the demand for frequencies of each WRAN and what is actually allocated. The algorithms are evaluated and it is concluded that they can efficiently assign channels to WRAN taking into account the bandwidth needs of each network. [ABSTRACT FROM AUTHOR]

We propose an efficient algorithm based on continuous-time Markov chain to compute the availability of network connections protected by shared recovery schemes in optical networks. The shared schemes are widely used to improve the availability without excessively expending network resources. The basic idea consists to share the recovery resources between two or more connections in order to reduce the amount of extra resources. This approach assumes that only a limited number of simultaneous failures that overlaps may happen in the network. In the presence of overlapping failures that affect simultaneously many connections, some may not be recovered, determining a reduction on their availability. The order in which overlapping failures occur is important because the first affected connection will acquire the shared resources, preventing the remaining connections to activate the recovery. The algorithm proposed in this paper significantly simplifies the calculation of the steady-state vector. We show that our approach is computationally more efficient and more accurate than the traditional approach used in the literature. [ABSTRACT FROM AUTHOR]