Analysis of fault-tolerant routing algorithms in k-ary n-cube networks

Many exiting studies have focused on the design of efficient fault-tolerant routing algorithms for common multicomputer networks, such as k-ary n-cubes. However, there has been comparatively little research activity on analysing and comparing these algorithms. In an effort towards filling this gap, this paper analyses the performance of two fault-tolerant routing algorithms, namely the unsafety vectors and probability vectors, recently proposed for k-ary n-cubes. In the unsafety vectors algorithm, each node calculates numeric unsafety vectors capturing information about faulty nodes in the network and uses them to achieve efficient fault-tolerant routing. In the probability vectors algorithm, a probabilistic approach is employed to achieve fault-tolerance. Each node calculates a probability vector, where the ith element represents the probability that a destination node at distance i cannot be reached through a minimal path due to a faulty node or link. The results presented below reveal that both algorithms exhibit good performance characteristics for practical number of faulty nodes in the network. However, the probability vectors algorithm is superior over the unsafety vectors in terms of the computation and communication overhead.