No abstract

This work is supported by Makhina Robotics, Istanbul, Turkey. Ali Bolu worked as a robotics engineer for developing RMFS in Makhina Robotics.

In a non-geostationary satellite constellation with inter satellite links (ISLs), there could be many shortest paths between two satellites in terms of hop count. An efficient routing algorithm should effectively use these paths in order to distribute traffic to ISLs in a balanced way and to improve the performance of the system. This paper presents and evaluates a novel priority-based adaptive shortest path routing (PAR) scheme in order to achieve this goal. PAR sets the path towards the destination in a distributed manner, using a priority mechanism depending on the past utilization and buffering information of the ISLs. Moreover, to avoid unnecessary splitting of a flow and to achieve better utilization of ISLs, enhanced PAR (ePAR) scheme is proposed. This paper evaluates performance of the proposed techniques by employing an extensive set of simulations. Furthermore, since there are a number of ePAR parameters that should be adjusted depending on the network and traffic characteristics, a detailed analysis of ePAR scheme is provided to form a framework for setting the parameters. This paper also includes a method for adaptation of the proposed algorithms to minimum-delay path routing. been directed towards development of non-geostationary (NGEO) systems consisting of Low Earth Orbit (LEO) satellites and Medium Earth Orbit (MEO) satellites. For better utilization of satellites and increasing the performance of the system, new NGEO systems usually support intersatellite links (ISLs) between satellites. The use of ISLs raises the issue of routing in the satellite network. Employing an efficient routing protocol is a critical issue, since satellite network resources are costly and must be managed in an optimal way. Unfortunately, mobility of satellites complicates the routing issue in an NGEO satellite system; hence terrestrial routing algorithms are not directly applicable in satellite networks. Satellite movements cause both the dynamicity of network topology (variation in ISL lengths, etc.) and dynamicity of traffic passing over satellites. A good routing algorithm should be adaptive to these dynamics.It is important to note that changes in the network topology are periodic and predictable because of the strict orbital movements of satellites. Therefore, it is reasonable to use this periodicity feature while calculating routes. Dynamic Virtual Topology Routing (DVTR) [1] is one of the most common routing methods that use the periodicity of the topology changes. DVTR divides the system period into a set of time intervals, so that topology remains constant over each interval. Time intervals are chosen to be short enough to assume that costs of ISLs are fixed. Therefore optimal paths and alternate paths can be established using well-known methods like Dijkstra's shortest path algorithm. DVTR approach decreases the online computational complexity with the expense of large storage requirements.In a satellite network, routing decisions can be made offline or on-board. In the former case, the routing algorithm could ...

No abstract

No abstract