Abstract :
An analytical framework is presented to study the self-adaptive behavior of probabilistic routing protocols for computer networks. Such soft routing protocols have attracted attention for delivering packets more reliably, robustly, and efficiently than conventional deterministic approaches. Efficient global operating parameters can be estimated without resorting to expensive Monte-Carlo simulation of the whole system. Key model parameters are routing sensitivity and routing threshold/noise, which control the "randomness" of packet routes between source and destination, and a metric estimator. Global network characteristics are estimated, including steady state routing probabilities, average path length, and path robustness. The framework is based on a Markov chain analysis. Individual network nodes are represented as states. Standard techniques are used to find primary statistics of the steady state global routing pattern, given a set of link costs. The use of packets to collect information about, or "sample," the network for new path information is also reviewed. How the network sample rate influences performance is investigated.
Keywords :
Markov processes; computer networks; estimation theory; optimisation; probability; routing protocols; Markov chain analysis; Markovian termite; computer networks; global network characteristics estimation; metric estimation; packet routing; probabilistic routing protocols; self-adaptive behavior; soft routing protocols; steady state routing probabilities; Ad hoc networks; Computer network reliability; Costs; Noise robustness; Parameter estimation; Particle swarm optimization; Routing protocols; State estimation; Steady-state; Wireless application protocol;