Combating Bufferbloat in Multi-Bottleneck Networks: Theory and Algorithms

Bufferbloat is a phenomenon in computer networks where large router buffers are frequently filled up, resulting high queueing delay and variation. More more delay-sensitive applications on the Internet have made this pressing issue. Interacting with Transmission Control Protocol (TCP), active queue management (AQM) algorithms run routers play an important role combating bufferbloat. However, AQM not been widely deployed due to complicated manual parameter tuning. Moreover, they often designed analyzed based network models single bottleneck link, rendering their performance stability unclear multi-bottleneck networks. In paper, we propose general framework combat bufferbloat We first present equilibrium analysis for TCP/AQM system provide sufficient conditions uniqueness of point system. then decompose into single-bottleneck subsystems derive local asymptotic subsystems. Using our framework, develop algorithm compute further case study analyze recently proposed Controlled Delay (CoDel) devise Self-Tuning CoDel improve stability. Extensive numerical packet-level simulation results only verify theoretical studies but also show that significantly stabilizes networks, thereby mitigating