Cache-Aided Interactive Multiview Video Streaming in Small Cell Wireless Networks

TL;DR

This paper proposes a joint caching and scheduling strategy for interactive multiview video streaming in small cell wireless networks, optimizing video quality and delivery time, and introduces a greedy algorithm with provable approximation guarantees.

Contribution

It formulates a novel NP-hard joint caching and scheduling problem considering video quality and delay, and provides a greedy solution with a new approximation ratio for submodular maximization under knapsack constraints.

Findings

The proposed policy outperforms benchmark algorithms in simulations.

The greedy algorithm achieves a 1/2(1-e^{-1}) approximation ratio.

The approach effectively balances video quality and delivery delay.

Abstract

The emergence of novel interactive multimedia applications with high rate and low latency requirements has led to a drastic increase in the video data traffic over wireless cellular networks. Endowing the small base stations of a macro-cell with caches that can store some of the content is a promising technology to cope with the increasing pressure on the backhaul connections, and to reduce the delay for demanding video applications. In this work, delivery of an interactive multiview video to a set of wireless users is studied in an heterogeneous cellular network. Differently from existing works that focus on the optimization of the delivery delay and ignore the video characteristics, the caching and scheduling policies are jointly optimized, taking into account the quality of the delivered video and the video delivery time constraints. We formulate our joint caching and scheduling problem as the average expected video distortion minimization, and show that this problem is NP-hard. We then provide an equivalent formulation based on submodular set function maximization and propose a greedy solution with $\frac{1}{2}(1-\mbox{e}^{-1})$ approximation guarantee. The evaluation of the proposed joint caching and scheduling policy shows that it significantly outperforms benchmark algorithms based on popularity caching and independent scheduling. Another important contribution of this paper is a new constant approximation ratio for the greedy submodular set function maximization subject to a $d$-dimensional knapsack constraint.