A survey of backpressure based scheduling algorithms for delay tolerant networks

摘要

The Delay Tolerant Networks (DTN) are networks where the end-to-end paths between source and destination are unstable or unlikely. In such networks, conventional path-discovery-based MANET routing techniques like AODV and DSR are not possible because the network may not form a single connected partition at any time, and thus a full path may never exist between the source and the destination. In such cases a store and forward algorithm is required, where the data packets can be stored in the network and be delivered before its deadline is over. Backpressure is one such algorithm, which stores data packets in queues and tries to deliver data packets to their final destination by building up gradients. It has been proved that this algorithm works effectively during peak traffic period compared to other DTN algorithm such as epidermis, spray and wait, Maxprop, etc., This paper is brief study of backpressure algorithm from its origin to its developed versions at present. It also explains the methods that can improve the working of backpressure in high traffic as well as low traffic with reduced overloads at the nodes maintaining queues.