Controlling Congestion via In-Network Content Adaptation
Yongzhou Chen, Ammar Tahir, Radhika Mittal · 2022
Realizing that it is inherently difficult to match precisely the sending rates at the endhost with the available capacity on dynamic cellular links, we built a system, Octopus, that sends real-time data streams over cellular networks using an imprecise controller (that errs on the side of overestimating network capacity) and then drops appropriate packets in the cellular-network buffers to match the actual capacity. We designed parameterized primitives for implementing the packet-dropping logic, which the applications at the endhost can configure differently to express various content-adaptation policies. Octopus transport encodes the app-specified parameters in packet header fields, which the routers can parse to execute the desired dropping behavior. Our evaluation shows how real-time applications involving standard and volumetric videos can be designed to exploit Octopus for various requirements and achieve a performance that is 1.5 to 18 times better than state-of-the-art schemes.