In this paper, we propose a multi-access coded caching system under decentralized setting tailored for Content Delivery Networks (CDNs). In this system, a central server hosts N files, each of size F bits, and serves K≤N users through a shared link. The network is equipped with c caches, each with a capacity of MF bits, distributed across the network, where each of the K users is connected to a random set of r≤c caches. Initially, we consider a model where each cache subset is accessed by an equal number of users. We introduce a novel content delivery algorithm for the central server, which allows us to derive a closed-form expression for the per user transmission rate. Using techniques from index coding, we prove the optimality of the proposed delivery scheme. Additionally, we extend the model to propose a more general and novel framework by allowing each subset of caches to serve an arbitrary number of users, thereby greatly enhancing the system’s flexibility and applicability. We also propose a new delivery algorithm tailored to this generalized setting and demonstrate its optimality under specific user-to-cache association scenarios. Numerical results demonstrate that, in a specific scenario where the user-to-cache associations do not satisfy the optimality conditions, the proposed generalized scheme shows improvement over the order-optimal state-of-the-art decentralized multi-access coded caching scheme for small cache sizes. Specifically, when approximately 25% of the content is stored at every cache, the proposed scheme achieves up to a 20% reduction in the per user transmission rate. Considering that both schemes serve an equal number of users, the observed improvements indicate a potential reduction in server bandwidth requirements, lower latency, and enhanced energy efficiency during content delivery.

Caching has emerged as a potential way to reduce the latency of content delivery and decrease network traffic during peak hours. In this paper, decentralized caching is considered where caches are filled with random contents of the files. The shared caching problem is considered in which more than one user can access a cache. A precoding technique using erasure codes is employed on the files before the caching. It is shown that the precoding technique implemented improves the delivery rate as compared to the rate when no erasure precoding is employed. Moreover, it is established that the rate corresponding to the proposed decentralized scheme matches with that of the optimal centralized scheme for certain cache sizes. Hence for certain specific cache sizes, the proposed scheme is optimal.

Data traffic in a client-server framework exhibits a temporal variability leading to congestion of resources at peak hours. One prevalent technique to overcome this problem is to load popular content/data into cache memories distributed across the end users. In this paper, the multi-access coded caching problem is considered in which each client is connected to multiple consecutive caches in a cyclic wrap around fashion and the cache memories are arbitrarily loaded in a decentralized manner. A new delivery scheme is proposed for the decentralized multi-access coded caching problem. A lower bound on the delivery rate is also obtained for the decentralized multi-access coded caching problem using techniques from index coding. The delivery scheme is shown to be optimal among all linear schemes when the number of caches associated with each user satisfies certain constraints.

Coded caching is a technique which enables the server to reduce the peak traffic rate by making use of the caches available at each user. In the classical coded caching problem, a centralized server is connected to many users through an error free link. Each user have a dedicated cache memory. This paper considers the shared caching problem which is an extension of the coded caching problem in which each cache memory could be shared by more than one user. An existing prefetching and delivery scheme for the shared caching problem with better rate-memory tradeoff than the rest is studied and the optimality of the scheme is proved by using techniques from index coding. The worst case rate of the coded caching problem is also obtained by using cut-set bound techniques. An optimal linear error correcting delivery scheme is obtained for the shared caching problem satisfying certain conditions.