Information-Centric Networks: A Convex Joint Cache Allocation and Router Assignment

Information-Centric Networking (ICN) is an emerging communication paradigm that addresses the limitations of the current Internet architecture by shifting the focus from hostbased communication to content-based delivery. A key feature of ICN is in-network caching, which reduces redundant transmissions, access latency and improves scalability. In this paper, we investigate the joint problem of cache allocation and request routing in ICN, where caching and routing decisions are inherently interdependent. We formulate the total network cost as a function of both cache placement and content delivery paths, subject to cache capacity constraints and service feasibility requirements. The resulting optimization problem is non-convex. In order to solve it, we employ inner approximation techniques based on the Difference-of-Convex (DC) programming framework combined with convex relaxation and penalty methods. This results in an iterative algorithm that jointly optimizes cache configuration and routing assignments while ensuring convergence to near-binary feasible solutions. Extensive simulations, including Monte Carlo experiments over randomized topologies, demonstrate that the proposed framework significantly reduces transportation cost, converges within a small number of iterations, and scales efficiently with network size. Comparisons with baseline strategies such as Cache Everything Everywhere (CEE) and Probabilistic Caching (Probp) show that our method achieves near-optimal performance without incurring excessive redundancy, which further validates both the practicality and efficiency of the proposed optimization approach.