Throughput Optimization with Latency Constraints

Abstract

Modern datacenters are increasingly required to deal with latency-sensitive applications. A major question here is how to represent latency in desired objectives. Incorporation of multiple traffic characteristics (e.g., packet values and required processing requirements) significantly increases the complexity of buffer management policies. In this work, we consider weighted throughput optimization (total transmitted value) in the setting where every incoming packet is branded with intrinsic value, required processing, and slack (an offset from the arrival time when a packet should be transmitted), and the buffer is unbounded but effectively bounded by slacks. We introduce a number of algorithms based on priority queues and show that they are non-competitive; then we introduce a novel algorithm based on emulating a stack and prove a constant upper bound on its competitive ratio that tends to 3 as the slack-to-work ratio increases. We support our results with a comprehensive evaluation study on CAIDA network traces.