On Wireless Link Scheduling and Flow Control

TL;DR

This thesis advances wireless link scheduling and flow control by developing algorithms that improve spatial reuse, throughput, and delay performance in wireless mesh networks considering physical layer characteristics.

Contribution

It introduces new link scheduling algorithms based on graph models, analyzes a power-based splitting algorithm for random access, and explores flow control through an information-theoretic approach.

Findings

Proposed algorithms achieve higher spatial reuse with minimal complexity increase.

Splitting algorithm attains a maximum stable throughput of 0.5518 with lower delay.

Derived maximum entropy of flow considering covert information in packet lengths.

Abstract

This thesis focuses on link scheduling in wireless mesh networks by taking into account physical layer characteristics. The assumption made throughout is that a packet is received successfully only if the Signal to Interference and Noise Ratio (SINR) at the receiver exceeds the communication threshold. The thesis also discusses the complementary problem of flow control. (1) We consider various problems on centralized link scheduling in Spatial Time Division Multiple Access (STDMA) wireless mesh networks. We motivate the use of spatial reuse as performance metric and provide an explicit characterization of spatial reuse. We propose link scheduling algorithms based on certain graph models (communication graph, SINR graph) of the network. Our algorithms achieve higher spatial reuse than that of existing algorithms, with only a slight increase in computational complexity. (2) We investigate random access algorithms in wireless networks. We assume that the receiver is capable of power-based capture and propose a splitting algorithm that varies transmission powers of users on the basis of quaternary channel feedback. We model the algorithm dynamics by a Discrete Time Markov Chain and consequently show that its maximum stable throughput is 0.5518. Our algorithm achieves higher maximum stable throughput and significantly lower delay than the First Come First Serve (FCFS) splitting algorithm with uniform transmission power. (3) We consider the problem of flow control in packet networks from an information-theoretic perspective. We derive the maximum entropy of a flow which conforms to traffic constraints imposed by a generalized token bucket regulator (GTBR), by taking into account the covert information present in randomness of packet lengths.