Dynamic Control of RLC Buffer Size for Latency Minimization in Mobile RAN

End-to-end packet latency is a defining component of the Quality of Experience (QoE) delivered to end users by interactive network applications. The tight latency requirements of those applications may be violated when they share packet buffers with elastic applications that use greedy TCP sources for end-to-end transport. The accumulation of queuing delay by the elastic application degrades the latency for the interactive application, however light the throughput of the latter may be (as in the case of voice or online gaming). Buffer sharing is unavoidable in the RLC sublayer of the 3GPP RAN stack. To minimize its negative effect on interactive applications, we split the buffering between the RLC and PDCP sublayer. Then we equip the PDCP buffer with per-flow queues and apply to the RLC buffer a new dynamic sizing mechanism that enforces the lowest queuing delay that is compatible with the existing configuration of the RLC connection. Our dynamic sizing solution reduces the queuing delay of PING packets by up to 95% compared to the default configuration with RLC-only buffer, and up to 50% compared to a configuration with PDCP flow queues and static RLC buffer size.