Simulation-based Research On Fragmentation Strategies For Space Network DTN Protocols

Traditional TCP protocol is based on the hypothesis that there exists"permanent"end-to-end connection. However, due to the long latency and intermittent connectivity features in space networks, it's difficult or even impossible to create end-to-end TCP connections. To tackle this problem, the delay tolerant network (DTN) protocols should be applied, which add a new layer called "bundle layer" to provide hop-to-hop custody transfer mechanism. But the fragmentation mechanism is just briefly covered in DTN's bundle protocol, ignoring the intermittent connectivity and predictable link features in space network. In this thesis, on the basis of simulation of different size proactive fragmentation strategies on network performance, an proactive fragmentation was proposed.Firstly, the basic mechanism of bundle protocol and DTN protocol architecture implementation prototype were briefly reviewed, how the proactive fragmentation, reactive fragmentation and fragment reassemble were implemented in DTN boundle layer. Based on this, a new proactively fragment strategy was proposed, that can make full use of the scheduled link feature in space intermittent connectivity scenarios. By dynamically and proactively allocating sizes of data that is to be transferred according to access duration and transmit speed of each connection in source node , network performance, like the packet delivery ratio, can be improved.Finally, the intermittent connectivity scenario was analyzed in the STK through simulation. The scenario was constructed in the ONE software according to parameter settings in the STK. Then dtnd and ONE were connected to simulate under DTN architecture. Specifically, the four cases of no fragmentation, proactive fragmentation, reactive fragmentation and the proposed fragmentation were compared to study the impact of fragmentation on network performance. Results show that the proposed fragmentation can greatly raise the packet delivery ratio in intermittent connectivity scenarios.