Win-Coupon: An Incentive Framework for 3G Traffic Offloading Xuejun Zhuo1, Wei Gao2, Guohong Cao2, Yiqi Dai1 5/29/2016 1Tsinghua University, 2The Pennsylvania State University Outline Background and Motivation Existing work and Challenges Main Approach of Win-Coupon Evaluation Conclusion Background & Motivation 3G Data Traffic o The smart mobile devices offer ubiquitous Internet access and diverse multimedia authoring and playback capabilities. o 3G data traffic is growing at an unprecedented rate. • By 2014, an average broadband mobile user will consume 7GB of traffic per month. • The total data traffic throughout the world will reach 3.6 exabytes per month. Overload!!! Background & Motivation How to solve the 3G overload problem ? o Increase 3G network capacity 1. 2. Expensive Inefficient o Offloading part of the 3G data traffic to other networks •Delay Tolerant Networks - DTN is a general paradigm of network in which nodes communicate without stable and continuous end-to-end connectivity. • WiFi hotspots Low Cost Efficient Background & Motivation How to utilize DTNs or WiFi hotspots for 3G traffic offloading? Extra Delay •Short range •High mobility •Large data file Outline Background and Motivation Existing work and Challenges Main Approach of Win-Coupon Evaluation Conclusion Existing work and Challenges Offloading 3G traffic by DTNs o [1] WoWMoM 2011 designs a framework, called PUSH-and-TRACK, which includes multiple strategies to determine how many copies should be injected by 3G and to whom, and then leverages DTNs to offload 3G traffic. o [2] CHANTS 2010 provides three simple algorithms to exploit DTNs to facilitate data dissemination among mobile users, in order to reduce the overall 3G traffic. o [3] CHANTS 2008, [4] MOBIHOC 2009 propose social-based algorithms to improve the performance of data dissemination in DTNs. Offloading 3G traffic by WiFi hotspots o [5]MOBISYS 2010 measures the offloading potential of the public WiFi based on city wide vehicular traces. o [6]CoNEXT 2010 considers a more general mobile scenario, and present a quantitative study on the performance of on-the-spot and delayed offloading by using WiFi. However these works have not considered the satisfaction loss of the users when a longer delay is caused by traffic offloading! Existing work and Challenges How to stimulate users to leverage their delay tolerance for 3G traffic offloading? o Incentive mechanism The users with high delay tolerance and large offloading potential should be prioritized for 3G traffic offloading. How to retrieve users’ delay tolerance, and how to determine pricing model? o Auction bids 3G operator: buyer Mobile user: seller How to measure users’ offloading potential? o Stochastic analysis Outline Background and Motivation Existing work and Challenges Main Approach of Win-Coupon Evaluation Conclusion Main Approach of Win-Coupon Overview o Reserve-auction based incentive mechanism • Allocation: the 3G operator decides which bidders are the winners and how long they need to wait. • Pricing: the 3G operator decides how much to pay for each winner. o Optimal auction outcome • min. incentive cost • s.t. expected offloading traffic ≥ offloading target Main Approach of Win-Coupon User Delay Tolerance o Satisfaction function S(t) – represents the price that the user is willing to pay for the data service with the delay t. = 800 $ (Private value) Price: 1000 $ (Bid) Gain 200 $ (Utility) Reverse Auction OK. Deal! Pay 1000 $ (Market clearing price) o Involving a single buyer and multiple sellers, and the buyer decides its purchase based on the bids sent by the sellers. oBid (bi): It is submitted by bidder i to express i’s valuation on the resource for sale, which is not necessarily true. o Private value (xi): It is the true valuation made by bidder i for the resource. o Market clearing price (pi): It is the price actually paid by the buyer to bidder i. o Utility (ui): It is the residual worth of the sold resource for bidder i. Main Approach of Win-Coupon Bidding Main Approach of Win-Coupon Bidding o tbound : Upper bound of user’s delay tolerance Resources for sale o User can divide tbound into multiple time units to sell o bid format: b = {b1, b2, …, bl} Main Approach of Win-Coupon Allocation Main Approach of Win-Coupon Allocation o Goals: • Which bidders are the winners • How long they need to wait o Problem formulation • min. incentive cost • s.t. expected offloading traffic ≥ offloading target NP-Hard!!! Main Approach of Win-Coupon Pricing Main Approach of Win-Coupon Pricing o Goals: • How much to pay for each winner • Assure Truthfulness and Individual Rationality o VCG-style pricing • Each winning bidder is paid by its “opportunity cost” • Opportunity cost: the cost of an alternative that must be forgone in order to pursue a certain action – the benefits you could received by taking an alternative action. •3G operator pays bidder i the coupon with value equal to the “opportunity cost” exerted to all the other bidders due to i’s presence. c1: The total value of coupons requested by all the bidders under the optimal allocation solution without the presence of bidder i c2: The total value of coupons requested by all the bidders except for bidder I under the current optimal allocation solution. Main Approach of Win-Coupon Pricing o Theorem 1: In Win-Coupon, for each bidder, say i, setting its bids truthfully, i.e., bi = xi, is a weakly dominant strategy. Truthfulness oTheorem 2: In Win-Coupon, all bidders are guaranteed to obtain non-negative utility. Individual Rationality Main Approach of Win-Coupon Predicting the bidders’ offloading potential Main Approach of Win-Coupon Predicting the bidders’ offloading potential o Network Model • Contact frequency: Exponential inter-contact time • Short contact duration: Random Linear Network Coding (RLNC) o Stochastic Analysis • Ordinary Differential Equation (ODE) Outline Background and Motivation Existing work and Challenges Main Approach of Win-Coupon Evaluation Conclusion Evaluation Numerical Results o Verify the accuracy of our prediction model o Compared the numerical results derived by using Matlab to the actual results derived by using Monte-Carlo simulations. Accurate!! Evaluation Real trace driven simulations o UCSD trace: recording the contact distory of 275 HP Jornada PDAs carried by students over 77 days. o Synthetic user delay tolerance o Synthetic data query information o We generate 50 data items, and each contains 8 packets o Length of one auction round: one hour o Performance metrics: • Offloaded traffic • Allocated coupons • Average downloading delay Evaluation Impact of Num. of bidders Impact of Reserve Price 40% of the 3G traffic can be offloaded Evaluation Impact of User Delay Tolerance Evaluation Large-scale trace driven simulations o Large-scale trace includes 2750 nodes by replicating the nodes in UCSD trace 10 times o Length of one auction round: ten minute Simulation Results 40% of the 3G traffic can be offloaded The corresponding delay is less than 5 hours Outline Background and Motivation Existing work and Challenges Main Approach of Win-Coupon Evaluation Conclusion Conclusion We propose a novel incentive framework to motivate users to leverage their delay tolerance for 3G offloading. We design a reverse auction based incentive mechanism, Win- Coupon, and formally prove that it has two desirable properties: 1) Truthfulness, 2) Individual Rationality. Taking DTNs as a study case, we provide an accurate model to predict users’ offloading potential. We implement Win-Coupon by using realistic DTN trace UCSD, and evaluate its practical use. Thanks Q&A Network Security Lab, Computer Science and Technology, Tsinghua University Selected References [1] J. Whitbeck, Y. Lopez, J. Leguay, V. Conan, M. D, Amorim, “Relieving the wireless infrastructure: When opportunistic networks meet guaranteed delays”, in Proc. of IEEE WoWMoM, 2011. [2] B. Han, P. Hui, V. Kumar, M. V. Marathe, G. Pei, A. Srinivasan, “Cellular traffic offloading through opportunistic communications: A case study”, in Proc. of ACM CHANTS, 2010. [3] C. Boldrini, M. Conti, A. Passarella, “Modelling data dissemination in opportunistic networks”, in Proc. of ACM CHANTS, 2008. [4] W. Gao, Q. Li, B. Zhao, G. Cao, “Multicasting in delay tolerant networks: a social network perspective”, in Proc. of ACM MOBIHOC, 2009. [5] A. Bala, R. Mahajan, A. Venkataramani, “Augmenting mobile 3G using wifi”, in Proc. of ACM MOBISYS, 2010. [6] K. Lee, I. Rhee, J. Lee, S. Chong, Y. Yi, “Mobile data offloading: how much can wifi deliver?”, in Proc. of ACM CoNETX, 2010.