Network-Coding Approach for Information-Centric Networking

The current internet architecture is inefficient in fulfilling the demands of newly emerging internet applications. To address this issue, several over-the-top (OTT) application-level solutions have been employed, making the overall architecture very complex. Information-centric-networking (ICN) architecture has emerged as a promising alternative solution. The ICN architecture decouples the content from the host at the network level and supports the temporary storage of content in an in-network cache. Fundamentally, the ICN can be considered a multisource, multicast content-delivery solution. Because of the benefits of network coding in multicasting scenarios and proven benefits in distributed storage networks, the network coding is apt for the ICN architecture. In this study, we propose a solvable linear network-coding scheme for the ICN architecture. We also propose a practical implementation of the network-coding scheme for the ICN, particularly for the content-centric network (CCN) architecture, which is termed the coded CCN (CCCN). The performance results show that the network-coding scheme improves the performance of the CCN and significantly reduces the network traffic and average download delay.

💡 Research Summary

The paper addresses the growing inefficiencies of the current IP‑based Internet architecture in supporting emerging applications such as live video streaming, social networking, and IoT. While over‑the‑top (OTT) solutions like CDNs, web caching, and P2P have been layered on top of IP to improve performance, they increase system complexity. Information‑Centric Networking (ICN) is presented as a promising alternative that decouples content from hosts, introduces in‑network caching, and naturally creates a multisource, multicast delivery environment.

Recognizing that network coding has proven benefits in multicast and distributed storage scenarios, the authors investigate its applicability to ICN. They first establish, through rigorous graph‑theoretic analysis, that a solvable linear network‑coding scheme exists for ICN and can be found in polynomial time. The model uses a directed acyclic scale‑free graph G = (V,E) with unit‑capacity edges. By redefining the classic Min‑Cut Max‑Flow (MCMF) theorem for the ICN context, they prove (Theorem 1) that the number of edge‑disjoint paths between any set of publishers and a set of gateway caches equals the min‑cut, and that, over a sufficiently large finite field GF(q), intermediate nodes can linearly combine packets so that all receivers obtain the data simultaneously. Theorem 2 extends this result algebraically, showing that if coding vectors are chosen uniformly at random from GF(q) (with q large enough), the transformation matrices at each receiver are full‑rank with high probability, guaranteeing decodability via solving a linear system. The proof leverages the Schwartz‑Zippel lemma to bound the failure probability.

Building on this theoretical foundation, the authors propose a practical implementation for the Content‑Centric Network (CCN) architecture, termed Coded CCN (CCCN). The implementation preserves the existing Interest/Data exchange semantics of CCN. Data packets are augmented with a lightweight coding header containing the coding vector and a sequence identifier. When a cache router receives a segment, it may immediately generate a coded packet by linearly combining stored segments according to its locally chosen coefficients. The coded packet is then forwarded downstream. Crucially, the Pending Interest Table (PIT) remains functional: each coded Data entry is mapped to the corresponding pending Interests, ensuring that the name‑based routing and caching mechanisms of CCN are untouched. This design avoids the pitfalls of earlier proposals (e.g., NC‑CCN) that broke CCN naming rules or rendered the PIT ineffective.

The performance of CCCN is evaluated via extensive simulations across diverse topologies, request patterns, and cache‑replacement policies. Three key metrics are examined: (1) total network traffic (bits transmitted), (2) average download delay per content request, and (3) overall system throughput under congestion. Results show that CCCN reduces transmitted traffic by more than 30 % on average, because a single coded packet can satisfy multiple downstream receivers, eliminating redundant transmissions. Average download delay drops by 25‑40 %, especially in scenarios with low cache‑hit rates where traditional CCN would need to retrieve the same segment from distant publishers. Under high load, the coded multicast approach alleviates bottlenecks on shared links, leading to higher throughput compared to native CCN and IP‑based baselines.

The paper also situates its contribution relative to related work. Prior studies applied network coding to P2P, CDNs, or DONA, but few addressed practical ICN integration. The NC‑CCN proposal introduced coding but violated CCN’s naming semantics and disrupted PIT operation. L4C2 employed Random Linear Network Coding (RLNC) for low‑latency video streaming but assumed specific QoS models and did not target general content delivery. CCCN, by contrast, offers a generic, standards‑compatible coding layer that can be incrementally deployed in existing CCN deployments.

Finally, the authors acknowledge limitations and outline future research directions. The current scheme uses only linear coding; exploring more efficient codes such as Luby Transform (LT) or Raptor codes could further reduce overhead. Adaptive coding window management and retransmission strategies are needed for latency‑sensitive services. Security considerations—authentication and integrity verification of coded packets—remain open challenges. Integrating coding-aware cache replacement policies and extending the analysis to mobile or highly dynamic networks are also promising avenues.

In summary, the paper demonstrates that linear network coding is both theoretically sound and practically beneficial for ICN, delivering measurable gains in traffic efficiency, latency, and robustness while preserving the core principles of the CCN architecture.