Session 7: Performance and Topology |
|||
|
|||
Abstract: Clustering is a typical topology control method improving efficiency of wireless ad hoc networks. Many clustering schemes have been proposed in the literature, but no prior work has quantitatively compared the scalability of these protocols. We present the scalability analysis of clustering protocols, which include Lowest ID Clustering (LID), Least Cluster Head Change Clustering (LCC), Mobility Based Clustering (MOBIC), and ID Load Balancing Clustering (IDLBC). We also compare them with our Personal Network Clustering Protocol (PNCP) while explaining its features. Usually each clustering protocol focuses on different performance aspects such as low maintenance load, stable cluster structure, energy efficiency, and load balance, etc., trying to maximize a few aspects. We apply absolute scalability and relatively weak scalability analysis to these schemes. Our results indicate that all of the protocols are absolutely scalable with an increasing number of nodes and traffic in the ad hoc network. Further, PNCP is ‘more’ scalable than other protocols in the cluster maintenance phase. Pages: 188-195 |
|||
|
|||
Abstract: Topology characterization of wireless sensor networks is a key issue for network planning, management and control as well as for maximizing the network performance and for optimizing protocol design. In this paper we analyze the structure of the topologies formed by means of the IEEE 802.15.4 association procedure, i.e., trees rooted at sink nodes, and we verify how such structures affect the network performance. Specific attention is dedicated to understand how the network lifetime is affected by the type of topology formed during the network set-up. A comparison of topologies having a single sink and multisink ones is provided. The determination of performance in case of multi-sink scenarios is extremely important; nevertheless, to the best of our knowledge, this problem is tackled here for the first time as for IEEE 802.15.4. Pages: 196-203 |
|||
|
|||
Abstract: We study different variations of the random walk (RW) such as RW with memory, RW with lookahead, RW using highest degree and RW proportional to the degree for random graphs. One of our insights is that comparison of different RW strategies based on the expected hopcount or weight is not sufficient. The expected hopcount for certain RW variations such as RW using highest degree is small. However, these strategies generally lead to infinite loops. Furthermore, the simulations show that RWusing highest degree with look-ahead and memory is the most efficient algorithm for searching in random graphs. Pages: 204-211 |
|||
|
|||
Abstract: Ensuring seamless mobility for users is becoming one of the main objectives of ongoing research activities in the field of data communications. In this paper, we provide a low latency handoff performance for nested mobile network context by using fast handover mechanism. To allow this coexistence, we are based on an enhancing version of NEMO named Light-NEMO presented in our previous works. The Fast Light-NEMO solution proposed in this paper combines the two schemes (fast handoff and Light-NEMO) in order to provide a seamless mobility for devices in nested networks. The paper describes the new architecture and mechanisms and provides analytical and simulation results showing the new scheme performance. Pages: 212-219 |
Pages: 1 2
