Cache policies for linear utility maximization

Abstract

Cache policies to minimize the content retrieval cost have been studied through competitive analysis when the miss costs are additive and the sequence of content requests is arbitrary. More recently, a cache utility maximization problem has been introduced, where contents have stationary popularities and utilities are strictly concave in the hit rates. This paper bridges the two formulations, considering linear costs and content popularities. We show that minimizing the retrieval cost corresponds to solving an online knapsack problem, and we propose new dynamic policies inspired by simulated annealing, including DynqLRU, a variant of qLRU. For such policies we prove asymptotic convergence to the optimum under the characteristic time approximation. In a real scenario, popularities vary over time and their estimation is very difficult. DynqLRU does not require popularity estimation, and our realistic, trace-driven evaluation shows that it significantly outperforms state-of-the-art policies, with up to 45% cost reduction. Notice: This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible.

A Decomposition Framework for Optimal Edge-Cache Leasing

“…If they are inversely proportional, files of smaller size are preferred. Another work with similar objectives is [14]. b) FemtoCaching: The users in [4] are equivalent to the regions as defined in our work.…”

“…The equation to (14) comes from the development of the quadratic term and from omitting the additive constants which do not affect the optimal choice of values for the variables. In the following, we will denote the objective function of (14) by g(y).…”

“…However, their speed of convergence can be an issue. Here, we present a problem specific solution method that does not depend on convergence and solves (14) exactly and efficiently. Our method is based on the insight that the optimal solution lies within a one-dimensional subspace of its |Q|-dimensional domain.…”

“…The value f −1 q (y) represents the reference water level ν if y is the non-zero water volume in bucket q. Letŷ q := w q y q for all q ∈ Q. Then, the following problem is equivalent to (14): N q ), q ∈ Q. Note that N q > 0 for all q and thus the upper and the lower bound of y q cannot be fulfilled with equality simultaneously.…”

A Low-Complexity Approach to Distributed Cooperative Caching with Geographic Constraints

“…Dehghan et al [14] propose utilitydriven caching and develop online algorithms that can be used by service providers to implement various caching policies based on arbitrary utility functions. Neglia et al [35] show that even linear utilities help to cover quite a number of interesting particular cases of cache optimization. It is interesting to note that the authors of [35] have also used stochastic simulated annealing type algorithms for the solution of the cache utility optimization problem.…”