The ever-growing trend of mobile users to consume high quality videos in their devices has pushed the backhaul network to its limits. Caching at Small-cell Base Stations (SBSs) has been established as an effective mechanism for alleviating this burden. Next generation mobile networks promise high Access Network density, hence multiple SBS association options per mobile user may exist. In this paper, we study the joint problem of mobile user association to SBSs and content placement to SBS-collocated caches, aiming to further optimize the utilization of backhaul network. The problem is solved periodically, considering time intervals where the users' location to the system is assumed to be fixed. First, we decompose the joint problem into two sub-problems that are solve sequentially, namely the content preference similarity-driven user association and the demand aware content placement sub-problems. We then propose a heuristic that consists of multiple phases. In particular, at a preparation phase, it performs clustering of users based on the similarity of their content preferences, accounting also for geographical constraints. The resulting clusters are then utilized for the demand-aware association of users to the SBS, while the placement of content is driven by the resulting local demand in each SBS, and takes place at the end. We demonstrate the effectiveness of our heuristic by evaluating its performance against multiple schemes that either lack a preparation phase or do not account for geographical constraints. As it is evident through the numerical results, the user clustering that takes place during the preparation phase can increase the overall cache hit ratio up to 20%.

Caching at the edge of the radio network is increasingly viewed as a promising countermeasure to the staggering demand for mobile video content. The persistent orientation of newer generations of mobile communication systems towards lower latency and faster radio access speeds only strengthens the arguments in its favor. When content caching is coordinated with other radio resource management functions, in particular, the benefits for the end users and the network operator are significant. In this paper, we investigate these benefits in cache-enabled small cell networks that jointly control (i) the Small-Cell Base Stations (SBSs) that serve as network access points for the mobile users; and (ii) the content that is stored at the SBS co-located caches. Our main contribution is a fast-converging computationally simple heuristic algorithm that iterates between assigning users to small cells and content to SBS caches, to maximize the overall cache hit ratio. The algorithm solutions compete with the optimal assignments at small problem instances and outperform alternative solutions for larger instances, especially when the content demand exhibits spatial locality. Combining good performance with non-prohibitive complexity, the algorithm could become a valuable tool for small cell network operators seeking to optimize the use of radio network resources.

In this paper, we investigate the performance gains that are achievable when jointly controlling (i) in which Small-cell Base Stations (SBSs) mobile users are associated to, (ii) which content items are stored at SBS co-located caches and (iii) which content items are recommended to the mobile users who are associated to different SBSs. We first establish a framework for the joint user association, content caching and recommendations problem, by specifying a set of necessary conditions for all three component functions of the system. Then, we provide a concrete formulation of the joint problem when the objective is to maximize the total hit ratio over all caches. We analyze the problems that emerge as special cases of the joint problem, when one of the three functions is carried out independently, and use them to characterize its complexity. Finally, we propose a heuristic that tackles the joint problem. Proof-of-concept simulations demonstrate that even this simple heuristic outperforms an optimal algorithm that takes only caching and recommendation decisions into account and provide evidence of the achievable performance gains when decisions over all three functions are jointly optimized.