Parallel and distributed computation: numerical methods
Parallel and distributed computation: numerical methods
A Fast Distributed Shortest Path Algorithm for a Class of Hierarchically Clustered Data Networks
IEEE Transactions on Computers
Routing algorithms in communication networks
Routing algorithms in communication networks
A New HAD Algorithm for Optimal Routing of Hierarchically Structured Data Networks
IEEE Transactions on Parallel and Distributed Systems
The Design and Analysis of Computer Algorithms
The Design and Analysis of Computer Algorithms
Exact Convergence of a Parallel Textured Algorithm for Data Network Optimal Routing Problems
IEEE Transactions on Parallel and Distributed Systems
A study on the flow assignment efficiency of MENTOR algorithm
ICCOM'06 Proceedings of the 10th WSEAS international conference on Communications
Model of optimal paths design for GMPLS network and evaluation of solution
WSEAS TRANSACTIONS on SYSTEMS
Hi-index | 0.00 |
In the routing process to select the data paths for Hierarchically Aggregation/Disaggregation and Composition/Decomposition, (HAD) networks, a fast algorithm for finding optimum paths for dataflow is needed. In this research we propose an algorithm called the Reverse Shortest Path algorithm to improve the speed in the calculating procedure for finding the shortest paths. This algorithm performs the reversed calculation in stead of the forward calculation used in conventional algorithms. The demand in each original destination pair (OD pair) has been distributed to the sub OD pairs in each relevant subnetwork r(u,v) = r(u,l) = ... = r(l,k) = r(k,v) with l and k, the gateways and ancestors in the active path. For each different commodities, the parallel processing is carried out with the shared shortest path processing time of O(log(n)) which less than O(mlog(n)) of HAD algorithm[1] where, n is the number of nodes in the networks, M is the number of commodities in each cluster and m is a positive integer which is less than M. The proposed algorithms have been developed and tested on a simulated network of 200 nodes clustered into 20 groups. Each group uses a personal computer as the processor for the group. Ten data Monte Carlo simulation patterns were generated for the test. The first five patterns represent typical normal dataflows which largely consist of short distance communications. The other five patterns represent the worst case data communication scenario. Test results on the proposed Reverse Shortest Path algorithm show that, for the tested network, the algorithm has improves the speed in finding the shortest paths by 20% as compared to the conventional shortest path algorithm.