TIME DEPENDENT PRICING FOR LARGE-SCALE MOBILE NETWORKS OF URBAN ENVIRONMENT:FEASIBILITY AND ADAPTABILITY

 

ABSTRACT

Because of severe network congestion experienced during peak hours in the urban area, dynamic time-dependent pricinghas been proposed by some mobile operators to shift users’ data usage from peak hours to off-peak time slots. We look at theperformance of time-dependent pricing on a large scale cellular network comprising ten thousand base stations. Our investigationreveals two important observations. First, time-dependent pricing performs well in reducing the peak-average ratio of the overall trafficof the network. However, the single price used by the network does not achieve good performance when we look at base stations inspecific regions, such as office regions. Second, we observe that location is another important factor that affects the traffic profile of abase station. Therefore, location information should be considered for designing a pricing strategy as well. We propose a frameworkthat combines both spatial and temporal traffic patterns for data pricing. Our simulation on ten thousand base stations suggests thatour proposed scheme is able to achieve an average of 16% smaller peak-to-average ratio. With over 15% smaller peak-to-average ratioof more than half of base stations in office regions, the performance is 2_ better than that achieved by the state of the arttime-dependent data pricing systems.

EXISTING SYSTEM:

Cellular traffic patterns have been extensively investigatedfor understanding various perspectives of cellular networks.Cici et al. analyzed the relationship between the applicationinterests and mobility patterns based on 280, 000users of a 3G mobile network. Lee et al.  demonstratedthat the spatial distribution of the traffic density can be approximatedby a log-normal or Weibull distribution, whileWang et al. found that mobile traffic followed a trimodaldistribution on both spatial and temporal dimensions.Our previous work quantitatively characterizedthe phenomenon of traffic tide in a large city-scale mobilenetwork. In this work, through an analysis of large scalemobile traffic, we discover the interaction between urbanfunctional regions and traffic patterns of base stations. Basedon this observation, we combine both spatial and temporalinformation for determining the price of mobile data usage.Time-dependent pricing has been studied in electricityand transport systems for decades. Actually, it wasused in telephony networks for a long time. Inspired bythese pioneer research, time-dependent pricing is used bymany operators for operating cellular networks as well. Theuse of time-dependent pricing in cellular network can beclassified into two categories. The first type focuses on theoreticalanalysis. Jiang et al.presented a game-theoreticanalysis of how to balance the trade-off between profit maximizingfor an operator and social welfare maximizing foran unselfish “social planner”. Ghanem et al.focused ondecreasing peak loads and enhancing the network revenue.Batubara et al demonstrated that this kind of pricingschemes can maximize operators’ profit, as well as users’Grade-of-service (GoS).

PROPOSED SYSTEM:

Our key contributions are:_ We study the performance of time-dependent pricingon a large scale cellular network. Our investigationreveals that time-dependent pricing performs well inreducing the peak-average ratio of overall cellulartraffic within a network comprising ten thousandbase stations. Our benchmark on network capacityand operational cost reveals important insights aboutthe effectiveness of time-dependent data pricing._ We find that a single unified price used by a wholecellular network does not achieve good performancefor base stations in specific regions. This conclusioncomes from our investigation of geographical locationcontext embedded in traffic patterns of basestations. Five types of base stations, whose trafficpatterns are mapped to the resident, transport, office,entertainment, and non-specific regions, are identified.We find time-dependent data pricing performswell for base stations deployed in residential andentertainment regions. However, poor performanceis observed for base stations deployed in transportand office regions. Our further analysis reveals fundamentalfactors that contribute to this observation._ We propose a framework that is able to combinespatial context information and time for determiningthe data price. Our simulation suggests that theproposed model is able to achieve over 15% smallerpeak-to-average ratio of traffic for 50% base stationsin office regions, 2_ better than the performance ofpure time-dependent pricing system.

 

CONCLUSION

In this paper, we investigate the performance of timedependentpricing on a large scale cellular network deployedin an urban area. Our investigation reveals twoimportant discoveries. First, a single price used by the timedependentpricing system does not perform well for basestations deployed in specific locations, such as residential regions.Second, in addition to time, spatial information, suchas urban function regions, should be included in the designof a data pricing model. Inspired by the two observations,we propose a framework that is able to dynamically combineboth temporal and spatial information for determiningthe price of cellular data. Our simulation shows that we areable to reduce traffic peak-to-average ratio by an averageof 16%. When applied in office regions, the performance is2_ better than that of pure time-dependent pricing scheme.In our future work, we aim to improve this pricing schemeby considering users’ willingness to delay their usage whenthey can move between different locations. To fulfill this,not only the traffic and services information, but also theindividual mobility data of users are required.

REFERENCES

[1] Cisco Visual Networking Index, White Paper, Feb. 2015. availableonline at http://www.cisco.com/c/en/us/solutions/collateral/service-provider/visual-networking-index-vni/whitepaperc11-520862.pdf, 2015. Accessed 1 June 2015.

[2] T. H. Batubara, C. Y. Huat, and M. Singh. On modeling the effect ofpeak-load pricing mechanism to the telecommunication traffic. InVehicular Technology Conference (VTC 2010-Spring), 2010 IEEE 71st,pages 1–5. IEEE, 2010.

[3] M. Boiteux. Peak-load pricing. The Journal of Business, 33(2):157–179, 1960.

[4] S. Borenstein, M. Jaske, and A. Rosenfeld. Dynamic pricing,advanced metering, and demand response in electricity markets.Center for the Study of Energy Markets, 2002.

[5] C.-H. Chang, P. Lin, J. Zhang, and J.-Y. Jeng. Time dependent adaptivepricing for mobile internet access. In 2015 IEEE Conference onComputer Communications Workshops (INFOCOM WKSHPS), pages540–545. IEEE, 2015.

[6] B. Cici, M. Gjoka, A. Markopoulou, and C. T. Butts. On thedecomposition of cell phone activity patterns and their connectionwith urban ecology. In Proceedings of the 16th ACM InternationalSymposium on Mobile Ad Hoc Networking and Computing, pages 317–326. ACM, 2015.

[7] F. Corpet. Multiple sequence alignment with hierarchical clustering.Nucleic acids research, 16(22):10881–10890, 1988.

[8] I. CVX Research. CVX: Matlab software for disciplined convexprogramming, version 2.0. http://cvxr.com/cvx, Aug. 2012.

[9] J. Ding, Y. Li, and D. Jin. Characterizing the phenomenon oftraffic tide for large-scale mobile cellular data networks. InComputer Communications Workshops (INFOCOM WKSHPS), 2015IEEE Conference on, pages 45–46. IEEE, 2015.

[10] K. Ghanem, N. Z. Khan, and A. Mitschele-Thiel. Peak load reductionon the mobile networks by applying new pricing policies.In 2007 4th International Symposium on Wireless CommunicationSystems, pages 446–450. IEEE, 2007.