A Hybrid Loss-Delay Gradient Congestion Control Algorithm for the Internet

Congestion Control (CC) has a significant influence on the performance of Transmission Control Protocol (TCP) and the Internet in general. In large bottleneck buffers, the standard TCP CC capacity probing strategy harms interactive and latency-sensitive applications, leading to a poor quality of experience. Delay-based CC reacts to congestion before queuing delay reaches a high value. In this thesis, we significantly enhance the CAIA-Delay Gradient (CDG) algorithm to create Hybrid Loss-Delay Gradient (HLDG), a sender-side host-to-host CC for the Internet to provides higher performance, lower queuing delay and better coexistence with TCP flows.