Improved Throughput Bounds for Interference-Aware Routing in Wireless Networks

We propose new algorithms and improved bounds for interference-aware routing in wireless networks. First, we prove that n arbitrarily matched source-destinations pairs with average distance d, for any 1 ≤ d ≤ √ n, in an O(n) size grid network achieve throughput capacity Ω(n/d). By a simple packing argument, this is also an upper bound in the worstcase. We show that, interestingly, the Ω(n/d) throughput can be achieved with single path routing, and present a simple distributed algorithm to compute these routes. For arbitrary networks, we propose a new node-based linear-programming (LP) formulation that leads to an improved worst-case throughput bound. Specifically, we show that the throughput delivered by our algorithm is at least 1/3 of the optimal, improving the previous best of 1/8. In addition, we show that for certain special topologies, such as tree-structured networks, our linear program yields optimal throughput. The LP-based methods split flows along multiple paths,...

• Type: Conference
• Year: 2007
• Where: COCOON
• Authors: Chiranjeeb Buragohain, Subhash Suri, Csaba D. Tóth, Yunhong Zhou
• Keywords: Achieve Throughput Capacity, COCOON 2007, Single Path, Throughput