Efficient Content Delivery through Provider Accessible Storage Service

Wireless data consumption has been growing at phenomenal rates over the past few years, resulting in tremendous pressure for service providers to accommodate what appears to be an insatiable data need from mobile subscribers. Unlike wired networks, the spectrum for wireless communication is finite and capacity expansion often requires massive cost for bandwidth acquisition and infrastructure setup. Hence, how to preserve on-going infrastructure investments without simply falling back to acquiring more spectrum has emerged as a significant challenge for the wireless industry. Solutions have been proposed along two primary tracks, namely instrumenting smaller areas with small cells for more capacity and / or mollifying the effect of data tsunami through better content management.

Although efforts towards small cells have demonstrated reasonable promise, a variety of technical challenges exist with regards to complexity, managements, and even economic viability. While approaches across the content management space offer opportunities to reduce the impact of burgeoning wireless traffic, nearly all of the techniques still operate on the primary wireless channel which limits their efficacy. We believe a critical opportunity for mobile devices has been overlooked from existing works and is ripe for exploration, namely the fact that modern devices have reasonably adequate free storage space which can be potentially utilized by the content provider for the purpose of content pre-staging.

To that end, my research efforts are focused on how to leverage the latent storage space (Provider Accessible Storage Service, i.e., PASS) on mobile devices to enable content pre-staging for reducing peak network traffic. The notion of PASS is to allow low-power, reasonable bandwidth transmission of data through secondary links to either directly share contents others have found popular or to signal content popularity for efficient pushing from upstream via broadcast, therefore bringing new methods to flatten or mollify the demand curve. This dissertation presents my research on investigating the potential for the said mechanism and discusses in detail how the proposed method has been implemented to show new opportunities for efficient content delivery.