Coded Caching with Heterogeneous File Demand Sets — The Insufficiency of Selfish Coded Caching

This paper investigates the integration of coded caching (CC) into fluid antenna (FA) multiple access (FAMA) systems to overcome their fundamental performance limitations. While FAMA has demonstrated strong interference suppression capabilities without relying on precoding, its delivery rate saturates in the high signal-to-noise ratio (SNR) regime due to residual multi-user interference, and it requires at least as many transmit antennas as served users. On the other hand, CC eliminates multi-user interference by jointly designing caching and delivery phases, but suffers from the well-known worst-user bottleneck, especially under wireless fading channels and low SNR conditions. To overcome the interference-limited nature of FAMA and the worst-user bottleneck in CC, we consider a CC-enabled FAMA framework. The proposed approach enables interference-free transmission to multiple users using a single transmit antenna, and leverages adaptive port selection at the users to combat channel fading. We analyze the average rates and effective gains of XOR-based CC and the recently developed aggregated CC (ACC) schemes under the FAMA framework, and derive simple closed-form approximations that accurately characterize the performance. Furthermore, we rigorously prove that, in the limit where the number of FA ports grows without bound while maintaining sufficient spatial diversity, the effective multiplexing gains in the low-SNR limit of both XCC- and ACC-enabled FAMA asymptotically achieve the nominal gain that is only attainable under high-SNR conditions with traditional antennas. Simulation results show that the proposed architecture outperforms both conventional FAMA and traditional-antenna CC schemes, achieving significant spectral efficiency gains in a wide range of SNR regimes.

The coded caching problem with shared caches where the server is equipped with multiple transmit antennas is considered. In a shared cache network, several users share a cache, but each user can access only a single cache. The fundamental limits of coded caching are known for the above setting under centralized uncoded placement. In the centralized case, to achieve the gains offered by coded caching, one requires a sub-packetization level which increases exponentially with the number of caches. The dedicated cache networks too had a similar issue, and placement delivery arrays (PDAs) were introduced as a solution to address the sub-packetization bottleneck of coded caching in the dedicated cache network. Our objective is to design coded caching schemes for shared caches with lower sub-packetization requirements than the existing schemes. By leveraging the PDA constructions, we propose a procedure to obtain new coded caching schemes for shared caches with reduced sub-packetization levels than the known schemes in both single and multi-antenna settings. The advantage of the proposed procedure is that all the existing PDA constructions can be transformed into coded caching schemes for shared caches, thereby reducing the sub-packetization level requirement. In addition, the proposed procedure also recovers the optimal caching schemes given by Parinello <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">et al.</i> by using a PDA corresponding to the Maddah-Ali Niesen scheme.

Coded caching was introduced as a technique of systematically exploiting locally available storage at the clients to increase the channel throughput via coded transmissions. Most known coded caching schemes in literature enable large gains in terms of the rate, however at the cost of subpacketization that is exponential in $K^{\frac{1}{r}}$ ($K$ being the number of clients, $r$ some positive integer). Building upon recent prior work for coded caching design via line graphs and finite-field projective geometries, we present a new scheme in this work which achieves a subexponential (in $K$) subpacketization of $q^{O((\log_{q}{K})^2)}$ and rate $\Theta\left(\frac{K}{(\log_{q} {K})^2}\right)$, for large $K$, and the cached fraction $\frac{M}{N}$ being upper bounded by a constant $\frac{2}{q^{\alpha-1}}$ (for some prime power $q$ and constant $\alpha \geq 2$) . Apart from this asymptotic improvement, we show that through some numerical comparisons that our present scheme has much lower subpacketization than previous comparable schemes, however with an increased rate and some increase in cache size required. For instance, we obtain practically relevant subpacketization levels such as $10^2 - 10^7$ for $10^2 - 10^4$ number of clients. Leveraging prior results on adapting coded caching schemes for the error-free broadcast channel to device to device (D2D) networks, we obtain a low- subpacketization scheme for D2D networks also, and give numerical comparison for the same with prior work.

Coded Caching is an efficient technique to reduce peak-hour network traffic. One limitation of known coded caching schemes is that the demands of all users are revealed to their peers in the delivery phase. Schemes that assure privacy for user demands are studied in the recent past. Assuming that the users are equipped with caches of small memory sizes, the achievable rate under demand privacy constraints is investigated in this work. We present an MDS code based demand private coded caching scheme with <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$K$</tex> users and <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$N$</tex> files that achieves a memory rate pair <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\left(\frac{1}{K(N-1)+1}, N\, \left(1- \frac{1}{K(N-1)+1}\right)\right)$</tex> . The presented memory-rate pair meets the lower bound under demand-privacy requirements, proposed by Yan and Tuninetti in the recent work “Fundamental Limits of Caching for Demand Privacy against Colluding Users”. By memory sharing, the achievable memory-rate pair characterizes the exact rate-memory trade-off for the demand private coded caching scheme for cache memory <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$M\in\left[0,\frac{1}{K(N-1)+1}\right]$</tex> .

In the coded caching problem apart from transmission rate, reducing the number of packets $F$ ia also meaningful for practical implementation for coded caching scheme. Yan et al. reformulate the centralized coded caching schemes as designing a corresponding placement delivery array (PDA). In this letter, we propose two concatenating constructions by concatenating some known PDAs to obtain new PDAs therefore new code caching schemes with low sub-packetization level. Further, the comparison and analysis of parameters between our new schemes and known schemes is also performed.

This paper studies device to device (D2D) coded-caching with information theoretic security guarantees. A broadcast network consisting of a server, which has a library of files, and end users equipped with cache memories, is considered. Information theoretic security guarantees for confidentiality are imposed upon the files. The server populates the end user caches, after which D2D communications enable the delivery of the requested files. Accordingly, we require that a user must not have access to files it did not request, i.e., secure caching. First, a centralized coded caching scheme is provided by jointly optimizing the cache placement and delivery policies. Next, a decentralized coded caching scheme is developed that does not require the knowledge of the number of active users during the caching phase. Both schemes utilize non-perfect secret sharing and one-time pad keying, to guarantee secure caching. Furthermore, the proposed schemes provide secure delivery as a side benefit, i.e., any external entity which overhears the transmitted signals during the delivery phase cannot obtain any information about the database files. The proposed schemes provide the achievable upper bound on the minimum delivery sum rate. Lower bounds on the required transmission sum rate are also derived using cut-set arguments indicating the multiplicative gap between the lower and upper bounds. Numerical results indicate that the gap vanishes with increasing memory size. Overall, the work demonstrates the effectiveness of D2D communications in cache-aided systems even when confidentiality constraints are imposed at the participating nodes and against external eavesdroppers.

For the caching problem, when the number of files is no larger than that of users, the best known rate-memory region is achieved by memory sharing between the rate-memory pairs obtained by three schemes: the scheme proposed by Yu et al. , the scheme proposed by Gomez-Vilardebo, and the scheme proposed by Tian and Chen. While the first two schemes operate on the binary field, the Tian–Chen scheme makes use of a finite field of order $2^{m}$ with, in some situations, $m\geq K\log _{2}(N)$ for a caching systems with $K$ users and $N$ files. The practical implications of this increase in the size of the field are equivalent to an increase, by a factor of $m$ , in the number of subfile partitions required. We propose a novel caching scheme that approaches the rate-memory region achieved by the Tian–Chen scheme as the number of users in the system increases, which only requires a field of order $2^{2}$ .

We consider a network consisting of a file server connected through a shared link to a number of users, each equipped with a cache. Knowing the popularity distribution of the files, the goal is to optimally populate the caches, such as to minimize the expected load of the shared link. For a single cache, it is well known that storing the most popular files is optimal in this setting. However, we show here that this is no longer the case for multiple caches. Indeed, caching only the most popular files can be highly suboptimal. Instead, a fundamentally different approach is needed, in which the cache contents are used as side information for coded communication over the shared link. We propose such a coded caching scheme and prove that it is close to optimal.

We consider a network consisting of a file server connected through a shared link to a number of users, each equipped with a cache. Knowing the popularity distribution of the files in the database, the goal is to optimally populate the caches such as to minimize the expected load of the shared link. For a single user, it is well known that caching the most popular files is optimal in this setting. However we show here that this is no longer the case for multiple users. Indeed, caching only the most popular files can be highly suboptimal. Instead, a fundamentally different approach is needed, in which the cache contents are used as side information for coded communication over the shared link. We propose such a coded caching scheme and prove that it is close to optimal.

As one of essential techniques to support the multimedia transmissions over the fifth generation (5G) mobile wireless networks, the statistical quality of service (QoS) provisioning has been proved to be effective in statistically guaranteeing time sensitive video transmission over time-varying wireless channels. But, how to efficiently design and implement statistical QoS provisioning enabled wireless network remains as the open challenging problem. To overcome these challenges, in this paper we propose the hierarchical-caching architecture to efficiently implement the statistical QoS guarantees in 5G wireless networks, which cooperatively caches the popular multimedia data contents at different wireless network caching tiers, efficiently trading the widely available storage resources for precious wireless bandwidth resources. However, the hierarchical caching scheme imposes the multi-hop wireless link statistical delay-bounded QoS provisioning problem when the requested multimedia data contents are fetched across tiers. In particular, we develop two hierarchical caching schemes to adaptively cache the delay bounded multimedia data contents: (1) caching in mobile users, which yields the single-hop wireless transmission and (2) caching in routers, which yields the tandem two-hop wireless transmission. To maximize the average effective capacity when cached data contents dynamically switch between these two schemes, we apply the Markov chain model to derive the optimal departure rates for data contents leaving cacheable mobile users and routers, respectively. U sing these two departure rates, we develop the optimal caching schemes in terms of the probability distribution of how long to cache the data contents and the probability of where to cache the data contents. Finally, we evaluate and validate these two optimized caching schemes through numerical analyses.

For ultra-dense networks with wireless backhaul, caching strategy at small base stations (SBSs), usually with limited storage, is critical to meet massive high data rate requests. Since the content popularity profile varies with time in an unknown way, we exploit reinforcement learning (RL) to design a cooperative caching strategy with maximum-distance separable (MDS) coding. We model the MDS coding based cooperative caching as a Markov decision process to capture the popularity dynamics and maximize the long-term expected cumulative traffic load served directly by the SBSs without accessing the macro base station. For the formulated problem, we first find the optimal solution for a small-scale system by embedding the cooperative MDS coding into Q-learning. To cope with the large-scale case, we approximate the state-action value function heuristically. The approximated function includes only a small number of learnable parameters and enables us to propose a fast and efficient action-selection approach, which dramatically reduces the complexity. Numerical results verify the optimality/near-optimality of the proposed RL based algorithms and show the superiority compared with the baseline schemes. They also exhibit good robustness to different environments.

The performance of cloud radio access networks (Cloud-RAN) is constrained by the limited fronthaul capacity. Physical layer caching is an effective technique to reduce the requirement on fronthaul capacity in peak time, which depends on the cache-hit ratio and the wireless transmission performance. Considering a dynamic points selection (DPS) based coordinated multi-point (CoMP) scheme to enhance the wireless transmission performance, this paper proposes an optimal probabilistic caching scheme to minimize the fronthaul capacity requirement in cloud-RAN. First of all, given the content caching probabilities, the average outage probability of cache-based services is derived. Based on the derivation, the relationship between the wireless transmission performance and the fronthaul capacity requirement is established. Then, an optimization problem is formulated to find an optimal probabilistic caching scheme to minimize the fronthaul capacity requirement, which can be solved using the interior point method. Simulation results show that using CoMP transmissions and optimizing the caching probabilities, the proposed scheme can reduce the fronthaul capacity requirement efficiently. Compared to the existing scheme which does not consider the effect of the wireless transmission performance on the fronthaul capacity requirement, the fronthaul capacity requirement can be reduced by 36\%. Moreover, the proposed scheme can also improve the cache service probability.

This paper considers the secretive coded caching problem with shared caches in which no user must have access to the files that it did not demand. In a shared cache network, the users are served by a smaller number of helper caches, and each user is connected to exactly one helper cache. To ensure the secrecy constraint in shared cache networks, each user is required to have an individual cache of at least unit file size. The existing secretive coded caching scheme for shared caches requires a subpacketization level, which is exponential in the number of helper caches. In this work, we propose a procedure to obtain new secretive coded caching schemes for shared caches with reduced subpacketization levels by utilizing the placement delivery array constructions. We also show that the existing secretive coded caching scheme for shared caches can be recovered using our procedure. In addition, a lower bound based on cut-set based arguments is derived for the shared cache networks under secrecy constraint and characterized the performance of the obtained scheme.

A device-to-device (D2D) aided multi-antenna coded caching scheme is proposed to improve the average delivery rate and reduce the downlink (DL) beamforming complexity. Novel beamforming and resource allocation schemes are proposed where local data exchange among nearby users is exploited. The transmission is split into two phases: local D2D content exchange and DL transmission. In the D2D phase, subsets of users are selected to share content with the adjacent users directly. In this regard, a low complexity D2D mode selection algorithm is proposed to find the appropriate set of users for the D2D phase with comparable performance to the optimal exhaustive search. During the DL phase, the base station multicasts the remaining data requested by all the users. We identify scenarios and conditions where D2D transmission can reduce the delivery time. Furthermore, we demonstrate how adding the new D2D phase to the DL-only scenario can significantly reduce the beamformer design complexity in the DL phase. The results further highlight that by partly delivering requested data in the D2D phase, the transmission rate can be boosted due to more efficient use of resources during the subsequent DL phase. As a result, the overall content delivery performance is greatly enhanced, especially in the finite signal-to-noise (SNR) regime.

Caching is one of the most important schemes for improving the performance of continuous media servers. Continuous object caching enables a server to support more clients simultaneously since it reduces the disk load imposed at each round. However, without a quantative analysis of the disk load reduction induced by caching, the caching effect can not be reflected in the admission control scheme, which limits the number of simultaneous clients serviced. In this paper, we define a performance metric for caching scheme in the continuous media server, define an optimal caching, formalize three heuristic block replacement model and propose a novel near optimal caching scheme. For quantative analysis of the proposed scheme we also propose a probabilistic model of the caching effect in a continuous media server. The proposed model enables the development of efficient statistical admission control algorithms that can increase the number of clients serviced simultaneously. To show the potential of the model, we present a simple example of a statistical admission control algorithm and demonstrate the performance enhancement resulting from the use of the proposed model.