A source‐driven probabilistic forwarding and caching strategy in NDN and SDN‐based NDN

Abstract

Summary NDN forwarding strategies are mainly developed receiver‐driven, where Interest forwarding and Data (content) delivery are controlled from the consumers' viewpoint. The receiver‐driven strategies are greedy by nature and emphasize optimizing the Interest forwarding, which is far from the performance bottleneck as an Interest is of tiny size in compared with a content's size. Thus, optimizing the content delivery can improve the forwarding performance. Moreover, source‐driven optimization of content delivery can alleviate the greedy nature of its counterpart. In this work, a source‐driven forwarding strategy is proposed where the content sources respond to Interest requests in a distributed manner. The content requests hit all the sources, but the sources respond probabilistically such that the content retrieval overhead and latency are minimized. The strategy consents the nearer sources to set a high probability of reacting with the matching content. On the contrary, the distant sources answer with an Offer to save transmission overhead. Besides, a caching strategy is proposed, an integral part of the source‐driven scheme, that caches diverse contents considering the gap between successive copies and content availability while admitting new contents. Next, we extend the source‐driven probabilistic forwarding strategy to operate in SDN‐based NDN architecture. The SDN‐based solution controls the content delivery phase using the global knowledge of NDN cache states. Simulation study conducted using ndnSIM reveals that the proposed forwarding and caching strategy improves user satisfaction and network performance in terms of different performance metrics such as transmission overhead, latency, and cache hit.