CTN Issue: May 2013

Today we are witnessing an unprecedented worldwide growth of mobile data traffic, which is expected to reach 10.8 exabytes/month in 2016, an 18-fold increase compared to 2011. These developments pose new challenges to mobile network operators (MNOs), who have to significantly enhance their infrastructure accordingly. However, traditional network expansion methods such as acquiring new spectrum licenses and upgrading technologies are costly and time-consuming, and more importantly are expected to be outpaced in less than 4 years by the continuing traffic increase. Clearly, operators must find novel methods to address this problem. Mobile data offloading through complementary network technologies, such as WiFi and femtocell, can significantly alleviate network congestion and enhance user QoS.

The paper envisions a data offloading marketplace managed by a centralized broker, where MNOs compete to lease third-party deployed WiFi or femtocell access points (APs) to offload the cellular traffic. Each MNO can employ multiple APs, and each AP can serve traffic multiple operators. However, AP owners are expected to ask for monetary compensations, since admitting MNOs' cellular traffic will consume APs' limited wireless resources and broadband connection capacities, and hence will affect APs' capabilities of serving their own internal traffic demand.

This paper proposes an double auction mechanism to coordinate the market. In the mechanism, MNOs declare how much they are willing to pay each AP. APs declare how much they need to get paid to offload traffic for each MNO. The broker collects the MNOs’ requests and the APs’ offers, and determines how much traffic of each MNO will be offloaded to each AP and at what price.

The challenges for designing an effective mechanism include:

1. MNO's offloading benefit is AP-specific. For example, an AP that is located at the boundary of an MNO’s cell is the most important one, since it can offload the traffic that otherwise would be very costly for the MNO to serve directly.
2. MNOs' offloading decisions are coupled. The accrued benefit from offloading a given amount of traffic to a certain AP depends on how loaded the MNO already is, which in turn depends on its decisions of offloading traffic to other APs.
3. The APs are heterogeneous. Different APs may have different costs for serving cellular traffic from the same MNO. The same AP may also have different costs for serving traffic from different MNOs due to different user and operator requirements.
4. APs' offloading decisions are coupled. The AP’s cost for offloading certain amount of data for one MNO also depends on the total traffic that the AP has committed to offload for other MNOs.

The proposed scheme in this paper resolves the above challenges, maximizes the social welfare (efficiency), guarantees that the broker does not lose money (weakly budget balanced), make MNOs and APs willing to participate (individually rational), and enable MNOs and APs reveal their truthful needs/demands (incentive compatible).

This paper wins a Best Paper Award in WiOPT 2013 conference.

Title and author(s) of the original paper in IEEE Xplore:
Title: An Iterative Double Auction Mechanism for Mobile Data Offloading
Author: George Iosifidis, Lin Gao, Jianwei Huang and Leandros Tassiulas
This paper appears in: 11th International Symposium on Modeling and Optimization in Mobile, Ad Hoc, and Wireless Networks (WiOPT)
Issue Date: May 2013

Statements and opinions given in a work published by the IEEE or the IEEE Communications Society are the expressions of the author(s). Responsibility for the content of published articles rests upon the authors(s), not IEEE nor the IEEE Communications Society.