The Network Neutrality Debate: An Engineering Perspective

Transcription

1 The Network Neutrality Debate: An Engineering Perspective Vishal Misra Columbia University, in the City of New York Joint work with Richard (Tianbai) Ma, Dahming Chiu, John Lui and Dan Rubenstein

2 Conversation between a prominent Economist and Dave Clark (Foundational Architect of the Internet) Economist: The Internet is about routing money. Routing packets is a side-effect. Economist: You really screwed up the money-routing protocols. Dave: We did not design any money-routing protocols. Economist: That s what I said.

3 Rest of the talk Background Cooperative Games and Shapley Values Application of Shapley Values to Peering Instability of settlement free peering Analysis of Paid Prioritization Monopoly Oligopoly Public Option

4 ISP Settlements and Shapley Values

5 The P2P Battlefield: Engineering and Economics Proposed engineering approaches: ISPs: Drop P2P packets based on port number Users: Dynamic port selection ISPs: Deep packet inspection Users: Disguise by encryption ISPs: Behavioral analysis Comcast started throttling BitTorrent traffic It became evident to us the problem was rooted in Economics, not Engineering

6 What were the Economists saying?

7 Building blocks of the Internet: ISPs The Internet is operated by thousands of interconnected Internet Service Providers (ISPs). An ISP is an autonomous business entity. Provide Internet services. Common objective: to make profit. ISP ISP ISP

8 Three types of ISPs 1. Eyeball (local) ISPs: Provide Internet access to residential users. E.g. Time Warner Cable, Comcast, Verizon, AT&T 2. Content ISPs: Serves content providers E.g. Cogent, Akamai, Level3, Netflix (Content Distribution Networks) 3. Transit ISPs: Provide global connectivity, transit services for other ISPs. E.g. Tier 1 ISPs: Level3, AT&T, Telefonica, Tata C ISP T ISP B ISP

9 Cooperative Games Players: N Coalition: A Value: V(A) Coalition: B Value: V(B) Coalitions Value: V

10 Cooperative Game Theory Analyses coalition formation given value allocation Value allocation characterizes a solution of a game Some properties of interest in a solution Stability: Players do not want to deviate from the solution Fairness: Allocation to players reflects their contribution

11 Convex Games V is Convex if for all coalitions A, B, V(AUB)-V(B) V(A)-V(A B) Marginal contribution of a player increases with the size of the coalition it joins Natural model for networks Metcalfe s law V(n) = n 2 Odlyzko s law V(n) = n log n

12 Core and Shapley Value of Convex Games Unstable Solutions Shapley Value Stable Solutions (Core)

13 Stability of the Shapley value V({1}) = a, V({2}) = b V({1,2}) = c > a + b. Convex game: V(SUT)>= V(S)+V(T) Whole is bigger than the sum of parts.

14 Stability of the Shapley value V({1}) = a, V({2}) = b V({1,2}) = c > a + b. Convex game: V(SUT)>= V(S)+V(T) Whole is bigger than the sum of parts. Core: the set of efficient profit-share that no coalition can improve upon or block.

15 Stability of the Shapley value V({1}) = a, V({2}) = b V({1,2}) = c > a + b. Convex game: V(SUT)>= V(S)+V(T) Whole is bigger than the sum of parts. Core: the set of efficient profit-share that no coalition can improve upon or block. Shapley [1971] Core is a convex set. The value is located at the center of gravity of the core.

16 Axiomatic characterization of the Shapley value What is the Shapley value? A measure of one s contribution to different coalitions that it participates in. Shapley 1953 Myerson 1977 Young 1985 Shapley Value Efficiency Symmetry Dummy Additivity Efficiency Symmetry Fairness Efficiency Symmetry Strong Monotonicity

17 Efficiency, Symmetry Efficiency: Symmetry: All Identical Profit goes players to the get Players equal shares

18 Balanced Contribution (Fairness)

19 How do we share profit? -- the baseline case C 1 B 1 One content and one eyeball ISP Profit V = total revenue = content-side + eyeball-side Fair profit sharing: ϕ = ϕ = B1 C V

20 How do we share profit? -- two symmetric eyeball ISPs Axiomatic Solution: Symmetry: same profit for symmetric eyeball ISPs ϕ = ϕ = ϕ B 1 B 2 B Efficiency: summation of individual ISP profits equals V ϕ + 2ϕ = V C1 B Fairness: same mutual contribution for any pair of ISPs Unique solution (Shapley value) 1 ϕ V = ϕ 0 C 1 2 B 1 B 1 C 1 B 2 ϕ = 2 V C1 3 ϕ = B 1 V 6

21 How do we share profit? -- n symmetric eyeball ISPs B 1 C 1 B 2 B n Theorem: the Shapley profit sharing solution is 1 ϕ = V, ϕ = n B V n(n+1) C n+1

22 Results and implications of profit sharing 1 ϕ = V, ϕ = n B V n(n+1) C n+1 With more eyeball ISPs, the content ISP gets a larger profit share. Multiple eyeball ISPs provide redundancy The single content ISP has leverage. Content s profit with one less eyeball: The marginal profit loss of the content ISP: If an eyeball ISP leaves C The content ISP will lose 1/n 2 of its profit. If n=1, the content ISP will lose all its profit. ϕ = C Δϕ = n-1 V - n V = - 1 ϕ n n+1 n 2 C 1 B 1 C n-1 V n B n-1 B n

23 Profit share -- multiple eyeball and content ISPs C 1 B 1 C 2 B 2 C m B n Theorem: the Shapley profit sharing solution is m n ϕ = V, ϕ = V B n(n+m) C m(n+m)

24 Results and implications of ISP profit sharing m V ϕ B= n (n+m), ϕ C = n V m(n+m) C 1 B 1 Each ISP s profit share is Inversely proportional to the number of ISPs of the same type. Proportional to the number of ISPs of the other type. C 2 C m B 2 B n Intuition When more ISPs provide the same service, each of them obtains less bargaining power. When fewer ISPs provide the same service, each of them becomes more important.

25 Profit share -- eyeball, transit and content ISPs C 1 T 1 B 1 C 2 T 2 B 2 C m T k B n Theorem: the Shapley profit sharing solution is

26 Common ISP Business Practices: A Macroscopic View Two forms of bilateral settlements: T ISP Zero-Dollar Peering T ISP Provider ISPs $$$ Customer-Provider Settlement $$$ C ISP Customer ISPs B ISP

27 Achieving A Stable Solution: Theory v Practice Shapley Reality T ISP T ISP C ISP B ISP

28 Implications When CR BR, bilateral implementations: Customer-Provider settlements (Transit ISPs as providers) Zero-dollar Peering settlements (between Transit ISPs) Common settlements can achieve fair profit-share for ISPs. If CR >> BR, bilateral implementations: Reverse Customer-Provider (Transits compensate Eyeballs) Paid Peering (Content-side compensates eyeball-side) New settlements are needed to achieve fair profit-share. When Customer Side Competition << Content Side Competition Paid Peering Will Dominate

29

30

31 31

32 Paid Prioritization

33 Network Neutrality (NN) Better! Happy?

34 Paid Prioritization (PP) Happier?

35 Some earlier analysis of paid prioritization Treat the Internet as an M/M/1 Queue Queue Size = QoS

36

37 User-side: 3 Demand Factors

38 Demand Function Fractional Loss of Users vs Fractional Loss of Throughput

39 System Side: Rate Allocation

40 Uniqueness of Rate Equilibrium

41 ISP Paid Prioritization Capacity Charge Premium Class Ordinary Class

42 Monopolistic Analysis

43 Utilities (Surplus)

44 Type of Content

45 Monopolistic Analysis

46 Regulatory Implications

47 Oligopolistic Analysis

48 Duopolistic Analysis

49 Duopolistic Analysis: Results

50 Oligopolistic Analysis: Results

51 Regulatory Preference ISP market structure Monopoly Oligopoly User Utility

52

53 Wireless Net Neutrality

54 Differences in the wireless scenario Payment model is different Largely metered. Wired consumers largely unlimited Bandwidth a lot more scarce Even if we auction of all remaining spectrum (40%), it only covers one year of bandwidth growth (doubling every year) Zero rated (bandwidth usage doesn t count against caps) is a big problem Good guys in the wired world (Google, Facebook etc.) are promoting zero rated apps in the wireless world Last-mile is less of a problem. Creating competition easier Multi-homing via AppleSim a first step

55 References Richard T. B. Ma, Dah Ming Chiu, John C. S. Lui, Vishal Misra and Dan Rubenstein. On Cooperative Settlement Between Content, Transit and Eyeball Internet Service Providers. IEEE/ ACM Transactions on Networking, Volume 19, Issue 3, pp , June, Extended Version of CoNEXT 2009 paper. Featured in IEEE ComSoc Technology News, June 2014, Special Issue on Network Neutrality, the Internet & QoS. Richard T. B. Ma, Dah Ming Chiu, John C. S. Lui, Vishal Misra and Dan Rubenstein. Internet Economics: The Use of Shapley Value for ISP Settlement. IEEE/ACM Transactions on Networking, Volume 18, Issue 3, pp , June, Richard T. B. Ma and Vishal Misra. Congestion and Its Role in Network Equilibrium. IEEE Journal on Selected Areas in Communications, Volume 30, Issue 11, pp , December, Richard T. B. Ma and Vishal Misra. The Public Option: A Non-Regulatory Alternative to Network Neutrality. IEEE/ACM Transactions on Networking, Volume 21, Issue 6, pp , December, Featured in IEEE ComSoc Technology News, June 2014, Special Issue on Network Neutrality, the Internet & QoS. Richard T. B. Ma, John C. S. Lui and Vishal Misra. Evolution of the Internet Economic Ecosystem. IEEE/ACM Transactions on Networking, to appear.