No abstract

Abstract-Network coding is known to improve throughput by mixing information from different flows and conveying more information in each transmission. Recently some proposals have demonstrated the benefits of applying network coding to wireless networks with broadcast transmissions. It is expected that the opportunities for coding and the corresponding gains depend on the bit-rate chosen for determining routes and transmitting packets. However, the previous work on wireless network coding assumed a fixed rate and did not explicitly account for the interaction between rate selection and coding gain. In this paper, we define a new metric, expected coded time (ECT), that measures the total time needed by a node to deliver two packets to their receivers given the bit-rate for transmitting coded packets. We then investigate how the optimal bit-rate for coded packets differs from that for transmission of native packets individually. We also study the performance of network coding under different fixed bit-rates for the whole network. Our evaluation shows that 11 Mbps is the best default fixed rate for MIT Roofnet and 5.5 Mbps is mostly the optimal rate to transmit coded packets when the ideal individual bit-rate for each receiver is different.

Some of the same features that make MANETs attractive, such as mobility and self-organization, also lead to increased vulnerability to traffic analysis. Data on who is communicating with whom, how often, how much, and when is easily available to any eavesdropper within range of the wireless network. Even if the payload is encrypted, standard MANET protocols transmit enough header and routing information in the clear making traffic analysis relatively easy for attackers. But users of MANETs may want to resist traffic analysis for a variety of reasons, ranging from secrecy for government and industry to simple personal privacy for individuals. Traffic analysis is a threat to secure communication, either by identifying targets for attacks such as denial-of-service or encryption cracking, or by revealing communication relationships.

Abstract-Opportunistic routing (OR) schemes, such as ExOR, have been shown to provide significant throughput gains over traditional best-path routing schemes for wireless networks. Though the performance of OR schemes depend on the bitrate, they currently use a fixed rate for transmitting packets. While several schemes have been proposed for selecting bit-rate for unicast transmission to a single receiver, none of them are suitable for broadcast transmission to multiple receivers under OR. This paper attempts to maximize the benefits of OR with dynamic bit-rate selection. We first define a new metric, Expected Anypath Communication Time (ExACT), that captures the time to deliver a packet to destination with a given rate at each hop under OR. We then propose Bit-rate Selection for Opportunistic Routing (BitSOR) algorithm that minimizes ExACT for each pair of nodes in the network. We evaluate the performance of BitSOR using MIT Roofnet trace and demonstrate significant potential improvement with dynamic rate over OR with the best fixed rate.

Wireless sensor networks are a rising commodity that demonstrate merit in the research domain due to their versatile nature. While powerful and feature-laden, these systems also require careful configuration for optimal packet transmission. This experiment aims to examine these configurations. By varying clear channel assessment threshold (CCAT) values, it will be demonstrated through implementation that changing the thresholds influences overall efficiency. Conversely, it will be shown that mis-configuring a single node will only impact the network under certain conditions. This work may be parlayed into more elaborate energy conservation and transmission efficacy schemes.