Congestion Control Protocol for Named Data Networking

The original use of the Internet was to share expensive resources by addressing endpoints using IP addresses. These days the Internet is mostly used to access distributed content rather than to share resources, but we still end up addressing endpoints in order to access distributed content. Named Data Networking (NDN) is an approach where content names are addressable, rather than endpoints. NDN's main approach is moving the internet towards a content distribution architecture and provides better support to mobile devices compared to TCP/IP. One of the design concerns of NDN is congestion control, which is the topic of this paper. Many recent studies on congestion control in NDN suggest either using an end-to-end approach or a hop-by-hop approach. Our NDN Congestion Control Protocol (CCP) is designed to detect and control congestion at the consumer end without the use of RTT measurements or congestion signaling. CCP is a receiver driven, window based protocol that can detect and control congestion by making use of measurements of delay along the forwarding path. We implement CCP using ndnSim simulator and compare it to an end-to-end AIMDD (additive increase/multiplicative decrease) scheme that behaves the same way as TCP