Computer Science Department
School of Computer Science, Carnegie Mellon University
FCP: A Framework for an
Dongsu Han, Robert Grandl*,
Transport protocols must accommodate diverse application needs as well as network requirements. As a result, TCP has evolved over time with new congestion control algorithms such as support for generalized AIMD, background flows, and multipath. On the other hand, explicit congestion control algorithms have shown to be more efficient. However, they are inherently more rigid because they rely on in-network components. Therefore, it is not clear whether they can evolve to support diverse application needs at least as much as TCP. This paper presents a novel framework for network resource allocation that supports evolution by accommodating diversity and exposing a flexible but simple abstraction for resource allocation. The core idea is to design a resource allocation scheme that allows aggregation and local control. To implement this idea, we leverage ideas from economics-based congestion control (but without actual congestion charging) with explicit virtual price feedback. We show that our design, FCP, allows evolution by accommodating diversity and ensuring coexistence, while being as efficient as existing explicit congestion control algorithms.