Scheduling data transfers for Dynamic Pricing
22 July 2013
The rapid increase in data traffic is making mobile networks highly congested. Paradoxically, even such congested networks have an abundance of underutilized capacity due to significant inefficiencies created by uneven data usage. Solutions such as dynamic congestion based pricing can help tap this stranded capacity. Motivated by these trends we consider system for data transfers (e.g. content/app downloads, bulk uploads, backups) with discounts or credits given for delivery using network's spare capacity. We study the problem of scheduling such data transfers while maximizing the use of spare capacity. User mobility, temporal and spatial variations in spare capacity, transfer deadlines along with considerations for battery drain and signaling and data overheads makes this a challenging problem. We propose efficient algorithms for this problem and quantify their performance both analytically and via simulations. Our algorithms also have application in other context (transportation, utilities) where dynamic pricing can be effective for yield enhancement.
