The growing amount of traffic in mobile data networks is causing concern for Internet service providers (ISPs), especially smaller ISPs that need to lease expensive links to Tier 1 networks. Large amounts of traffic in “peak” hours are of especial concern, since network capacity must be provisioned to accommodate these peaks. In response, many ISPs have begun trying to influence user behavior with pricing. Timedependent pricing (TDP) can help reduce peaks, since it allows ISPs to charge higher prices during peak periods.We present results from the first TDP trial with a commercial ISP. In addition to analyzing applicationspecific mobile andWiFi traffic, we compare changes in user behavior due to monthly data caps and time-dependent prices. We find that monthly data caps tend to reduce usage, while TDP can increase usage as users consume more data during discounted times. Moreover, unlike data caps, TDP reduces the network’s peak-to-average usage ratio, lessening the need for network over-provisioning and increasing ISP profit.
|Original language||English (US)|
|Title of host publication||Passive and Active Measurement - 16th International Conference, PAM 2015, Proceedings|
|Editors||Yong Liu, Jelena Mirkovic|
|Number of pages||13|
|State||Published - 2015|
|Event||16th International Conference on Passive and Active Measurement, PAM 2015 - New York, United States|
Duration: Mar 19 2015 → Mar 20 2015
|Name||Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|
|Other||16th International Conference on Passive and Active Measurement, PAM 2015|
|Period||3/19/15 → 3/20/15|
Bibliographical noteFunding Information:
We gratefully acknowledge the assistance of our colleagues at the Matanuska Telephone Association. Part of the work was supported by NSF CNS-1117126.
© Springer International Publishing Switzerland 2015.