Congestion Prediction Based Load Balancing in IP/LEO Satellite Networks(Satellite Network)

概要

論文の詳細を見る
Low Earth Orbit (LEO) satellite networks have worldwide coverage with low power consumption and short propagation delay. These features facilitate the development of global mobile networks and are suitable for multimedia communications. Since the users of LEO satellite networks are not uniformly distributed due to the geography and the population dispersion, congestion may often occur in the certain satellites due to heavy traffic loads. As a solution to this issue, the traffic loads on the satellites covering the area with high user density should be distributed. In this paper, utilizing the fact that congestion level of a LEO satellite can be predicted on the basis of its orbit and from the notification from adjacent satellites, we propose a dynamic load balancing scheme based on congestion prediction. In the proposed method, each satellite detects areas where congestions tend to occur and inform others where these areas are. Satellites execute two types of load balancing mechanisms to repress occurrences of congestions in those areas. We evaluate the performance of the proposed scheme through several simulations. The simulation results demonstrate that the proposed method improve packet drop rate, throughput, and distribution of traffic.
社団法人電子情報通信学会の論文
2008-10-30