Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
Smart Data PricingSmart Data Pricing
Seong-Lyun Kim
slkim@yonsei ac [email protected]://web.yonsei.ac.kr/slkim
1
Contents• Data Pricing• Price Competition• Subsidyy• Spectrum Pricing (Auction)• Q&A• Q&A
2
Data Pricing
“Revenue for the system and use yresources more efficiently”
J Walrand Economic Models of Communication Networks NewJ. Walrand, Economic Models of Communication Networks. NewYork: Springer, Ch. 3, pp. 57-87, 2008.
3
Tragedy of Commonsg y• Each herdsman (user) imposes a total
t lit (i t f ) th th th texternality (interference) on the others that exceeds his own benefit increment.
G. Hardin, “The Tragedy of the Commons,” Science, 162: 1243-1248, 1968.
4
Classification
S. Sen, C. Joe-Wong, S. Ha, and M. Chiang. Smart data pricing (sdp): Economic solutions to network congestion. SIGCOMM eBook on Recent Advances in Networking, 2013.
5
Time-ahead Pricingg
S. Ha, S. Sen, C. Joe-Wong, Y. Im, and M. Chiang, "TUBE: Time Dependent Pricing for Mobile Data", Proc. of ACM SIGCOMM 2012
6
DataWiz
7
다른모습그러나비슷한속성
• 공통점• “Air”• 규모의경제 (Density of customers)에의존• 이윤추구하는동시에공공 서비스 (예 철도 전기 )• 이윤추구하는동시에공공 서비스 (예, 철도, 전기…)• 규제기관의강력한 통제
• 가격• 서비스• 시장진입
Y. J. Hwang, K. W. Sung, S.-L. Kim and J. Zander, “Scenario Making for Assessment of Secondary Spectrum Access,” IEEE Wireless Communications, Vol. 19 (4), pp. 25-31, August 2012.
88
Economic Models of CommunicationEconomic Models of Communication Networks
J. Walrand, Economic Models of Communication Networks. NewJ. Walrand, Economic Models of Communication Networks. NewYork: Springer, Ch. 3, pp. 57-87, 2008.
9
Revenue-max Pricingg• Price ≤ User type ≤ QoS
0 1
User type
10
Paris Metro Pricing (1/3)g ( )• Utility Depends on Utilization
– How much the service price should be?
A. ODLYZKO, “Paris Metro Pricing for the Internet,” ACM Conference on Electronic Commerce,1998.
11
Paris Metro Pricing (2/3)g ( )• Service differentiation to increase the revenue
substantiallysubstantially.
Inexpensive High Utilization Low QoS (tourist class)
• A more expensive network has a lower utilization and
Expensive Small Utilization High QoS (business class)
• A more expensive network has a lower utilization and, accordingly, lower delays and more throughput for each of its connections
12
of its connections
Paris Metro Pricing (3/3)g ( )• Service differentiation with Paris metro pricing
i th f 1/6 t 9/40 bincreasesvthe revenue from 1/6 to 9/40, or by 35%.
0 1
Tourist class Business class
J. Walrand, Economic Models of Communication Networks. NewJ. Walrand, Economic Models of Communication Networks. NewYork: Springer, Ch. 3, pp. 57-87, 2008.
13
Price Competition
S. M. Yu and S.-L. Kim, "Game-theoretic Understanding of Price Dynamics in Mobile Communication Services," to appear in IEEE Transactions on Wireless Communications, http://arxiv.org/abs/1304.3875
14
Economic Layery• Users and providers respond to economic
i ti d ff t th t kincentives and affect the network
NSPRegulation
NSPsRegulator
CournotCompetition (C it )
BertrandCompetition
(P i )
Demand
(Capacity) (Price)
Networks UsersQoS
15
User Type, Price and QoSyp ,
• Price ≤ User type ≤ QoSPrice User type QoS
0 1
<Perfectly segmented market>
min1 1p max
1 1q max2 2q min
2 2p
<Non-segmented market>
0 1
<Non segmented market>
min max1 2 2q min
2 2p max1 1q 1p
16
Two Users Non-cooperative Game:NSP’s Optimal StrategyNSP s Optimal Strategy
• Best response function
* 22 1
11 2 1If 0 3 3 2
pp p
*2 1 2
3 3 21 2 1If
3 2p p p
*2 1
1 1If 2 2
p p
17
NSP’s Optimal Strategyp gy• Best Response Function
2pBR1 1BR
2BR
1/ 2
2 / 3
1 / 3long jump
i
1p1/ 2 12 / 31 / 3
price war
18
Price War with Long Jumpg p
1pp
0 12 / 31/ 3 1/ 2
2p
0 12 / 31/ 3 1/ 2
J. Walrand, Economic Models of Communication Networks. NewYork: Springer, Ch. 3, pp. 57-87, 2008.
19
Regulation for Convergenceg g
• A regulation that limits the number of price level changes k th i l l t ilib i i tmakes the price levels converge to an equilibrium point
that is Pareto-optimal.
20
Subsidy
21
Two Subsidy Schemesy
• Price Subsidy– Giving price discount to users accessing network
service.
• Data (QoS) Subsidy– Providing a predefined amount of data (QoS) to usersProviding a predefined amount of data (QoS) to users
without any charge.
S M Yu and S L Kim "Guaranteeing User Welfare in Network Service:S. M. Yu and S.-L. Kim, "Guaranteeing User Welfare in Network Service: Comparison of Two Subsidy Schemes," Proc. ACM SIGMETRICS/Performance Workshop W-PIN (First Workshop on Pricing and Incentives in Networks), London, UK, 2012.
22
Price Subsidyy
S. M. Yu and S.-L. Kim, "Game-theoretic Understanding of Price Dynamics in Mobile Communication Services," to appear in IEEE Transactions on Wireless Communications, http://arxiv.org/abs/1304.3875
23
Data (QoS) Subsidy Scheme( ) y
• Data (QoS) Subsidy Scheme (Type 1)– The regulator gives all of the available spectrum
amount to mobile network operators (MNO) for free in return for providing a predefined amount of data (QoS) to users without any charge(QoS) to users without any charge.
24
Partial Spectrum Leasingp g
For businessFor free
• Data (QoS) Subsidy Scheme (Type 2)Dynamic control of the amount of leasing bandwidth ➡ “free services” to all– Dynamic control of the amount of leasing bandwidth ➡ free services to all users in a leftover bandwidth.
– MNO: Joint leasing and pricing decisions.Regulator: “Optimal spectrum price” for maximal profit and user welfare– Regulator: Optimal spectrum price for maximal profit and user welfare.
S. Y. Jung, S. M. Yu, and S.-L. Kim, "Utility-optimal Partial Spectrum Leasing for Future Wireless Networks,“ in Proc. IEEE Vehicular Technology Conference, Spring, Dresden, Germany, 2013 (Best paper award).
25
User Welfare and Profit Gain
26
Spectrum Pricing (Auction)
S. Y. Jung, S. M. Yu, and S.-L. Kim, "Asymmetric-valued Spectrum Auction and Competition in Broadband Wireless Services,” in Proc. WIOPT 2014. http://arxiv.org/pdf/1307.7838v1.pdf
27
LTE Spectrum Auction (2013.8)p ( )• Auction (ongoing this week)
• Carrier Aggregation vs. Wideband LTE (20 MHz Bandwidth, Rel 11)Ca e gg ega o s de a d ( 0 a d d , e )
28
Asymmetric–valued Spectrum Auctiony p
• Two MNOs compete in a first-price sealed-bid auction where two spectrum blocks A and B are auctioned off towhere two spectrum blocks A and B are auctioned off to them
• A and B are the same amount of spectrum (i e 10 MHzA and B are the same amount of spectrum (i.e., 10 MHz spectrum block).
29
Spiteful Behaviorp
아사다마오 김연아
30
Multi-stage Gamesg
31
Results• MNOs announce different equilibrium prices to
h idi th litusers, even when providing the same quality services.
• The market share leader, despite charging a higher price, may still dominate the game.
32
항공서비스산업은지금?
33
Q&A
34