On how to Mitigate the Packet Reordering Issue in the Explicit Load Balancing Scheme

Efficient load balancing is an important requirement for intelligent engineering of traffic over all-IP satellite networks. In this regard, the authors have recently proposed an explicit load balancing (ELB) scheme for next-generation LEO/MEO satellite systems. In ELB, a congesting satellite requests its neighboring satellites to forward a portion of data, originally destined to travel through the satellite, via alternative paths that do not involve the satellite. While this feature yields better traffic distribution and reduces the overall packet drops that may occur at the congesting satellite, it raises the so-called packet reordering issue. In connection-oriented protocols such as TCP, an out-of-order reception of packets generates duplicate acknowledgments that result in a gratuitous halving of the congestion window. This ultimately degrades the overall throughput of the network. To cope with packet reordering issue in ELB, we suggest some minor modifications to the TCP implementation at the receiver side to enable receivers to judge the actual reason beneath the out of order reception of packets. We compare the performance of our method to that of standard TCP and TCP-PR, a recently proposed scheme for persistent packet reordering. Depending on the traffic characteristics and the satellite constellation type, discussion on the advantages and pitfalls of each scheme is given.