Real-time centralized and decentralized out-of-order data transfer scheduling techniques

In the first part of this paper we analyze several centralized and decentralized decision making methods for the case of deadline-constrained real-time out-of-order data delivery. We perform a simulation study in which we implement several data transfer scheduling algorithms and compare their results. In the second part we propose a peer-to-peer architecture with mixed components (both centralized and distributed) for optimizing reliable out-of-order data transfers, but without deadline constraints. We also provide experimental results which validate the concepts behind the peer-to-peer architecture.