Binary decision diagram based generation algorithm for large scale graph distribution

This paper proposes an on the fly algorithm for graph distribution. The algorithm uses a new distributed method combined with the Binary Decision Diagram (BDD) for holding a global image of the system. Though, the proposed approach is a distributed algorithm, it uses little or none communication between the different nodes(sites). Hence it increases the fault tolerance of an unreliable network. In addition it preserves the workload balancing property. As a result, the proposed algorithm attempts to balance the workload between the different sites of the network and reduces as much as possible the inter-processors communications overhead result of reducing the inter-sites edges.