CRT: an Adaptive Routing Protocol for LEO Satellite Networks

Low-earth orbit (LEO) satellite networks which have a world wide coverage are becoming more and more important to provide Internet services. One of the technical challenges in using LEO constellations to route Internet traffic is designing appropriate and efficient network layer to be used on board. In particular, traditional IP routing seems to be too complex for a simple and static satellite topology. On the other hand, satellite routing protocols proposed so far do not react promptly to hot spots in traffic generation. In this paper, we propose CRT, a new adaptive routing protocol which dynamically adjusts routes to balance the traffic load over the whole constellation. The protocol is evaluated through simulation using different (realistic) traffic scenarios. Results show that the protocol reacts well to congestion in satellite constellation with enough satellites to have a reasonable number of alternative routes for a given message (at least 36 satellites). In particular, we show that the overhead due to control messages in CRT is very low, performance is competitive to non adaptive algorithms when traffic load is low, and end-to-end delay and drop ratio are greatly reduced using CRT in medium to high traffic load conditions