Evaluation of AQM schemes to support Low Latency in the Internet

Many modern applications require low latency transmission in the Internet which is not explicitly supported today. Most networks are optimized for high throughput and low loss. This increases latency due to high queuing delays. Only very few applications actually need both high thorughput and low latency, therefore the network could offer a new parallel Internet service that provides lower latency compared to the best effort service that is usually offered today. This can be achieved by separating flows into different queues at the bottleneck link that operate either independently or could be coupled as propoded by the DualQ Coupled AQM for Low Latency, Low Loss and Scalable Throughput (L4S) [1].