PCach: The Case for Pre-Caching your Mobile Data

TL;DR

PCach proposes a smartphone-based pre-caching system using WiFi to reduce cellular network congestion by predicting and caching data, based on extensive user data analysis.

Contribution

It introduces a novel pre-caching algorithm that predicts optimal content to cache on smartphones, reducing cellular data use.

Findings

Median user transfers 15% of data via cellular network

80% of cellular data can be pre-cached via WiFi

Pre-caching algorithm predicts what and when to cache with high accuracy

Abstract

We present PCach, a smartphone-based approach for relieving the congestion in cellular network resulting from the exponential growth in mobile data traffic. The basic idea underlying PCach is simple: use WiFi to proactively cache content on the smartphone's memory, which otherwise would have been delivered through the cellular network. However, it leads to several challenging questions, including how much mobile data actually flows through cellular networks, how much data can be pre-cached, and when and what to pre-cache. We address these questions progressively using a thorough analysis of user data collected from our purpose-built crowdsensing Android application, actively utilized by 45 users for periods dating back to July 2014. Our analysis shows that the median smartphone user transfers 15% of their data via the cellular network and that 80\% of it can be pre-cached via WiFi. To capitalize on these results, we draw on a careful analysis of the measurement data to introduce an algorithm that can run stand-alone on off-the-shelf smartphones and predict with good accuracy when and what to pre-cache.