1. Soumya Sen Dept. of Electrical & Systems Engineering University of Pennsylvania ssoumya@seas.upenn.edu www.seas.upenn.edu/~ssoumya Joint Work with: R. Guerin, K. Hosanagar, Y. Jin 22 nd November, 2010. Georgia Institute of Technology. On the Adoption and Deployment of New Network Technologies: An Economic Perspective

2. Networked Systems have a ubiquitous presence –e.g., Internet, Power grid, Facilities Management networks, Distributed databases Success of new network technologies depends on: – Technical advantage – Economic factors (e.g. price, costs, demand) Shortcomings: –Many technologies have failed to get adopted e.g., IPv6 migration, QoS solutions –Ad-hoc decisions for deployment e.g. Cloud Computing, U-Verse versus. FiOS How to assess (design) new network technologies (architectures) for technical and economic viability? –Analytical frameworks –Multi-disciplinary approach S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Research Motivation 2

3. On Developing Analytical Models: –What are the key economic aspects? –What factors are specific to network technologies? (e.g. externality, gateways) –What are the ‘qualitative’ insights from the model? –Are the results robust to changes in the underlying model? Some Dimensions for Assessing Network Technologies: Topic 1: –Network Technology Adoption/ Migration How can a provider help its technology (service) to succeed? Topic 2: –Network Infrastructure Choice What infrastructure should the new technology (service) be deployed on? Understanding Trade-offs between Shared and Dedicated networks Topic 3: –Trade-offs between Functionality-rich versus Minimalist Designs S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Research Approach 3

4. Two competing and incompatible network technologies (e.g., IPv4 IPv6) –Different qualities and price –Different installed base Users individually (dis)adopt the technology that gives them the highest positive utility –Depends on technology’s intrinsic value and price –Depends on number of other users reachable (externality) Gateways offer a migration path –Overcome chicken-and-egg problem of first users Independently developed by each technology –Effectiveness depends on gateways (converters) characteristics/ performance Duplex vs. Simplex (independent in each direction or coupled) Asymmetric vs. Symmetric (performance/ functionality wise) Constrained vs. Unconstrained (performance/functionality wise) S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Problem Formulation 4

5. S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective A Basic User Model 5 Users evaluate relative benefits of each technology –Intrinsic value of the technology (  q 1 ) Tech. 2 better than Tech.1 (q 2 >q 1 )  denotes user valuation (captures heterogeneity), –Externalities: linear in no. of users (0≤x 1 +x 2 ≤1) - Metcalfe’s Law Possibly different across technologies (captured through β ) α i, 0  α i  1, i = 1,2, captures gateway’s performance –Cost (recurrent) for each technology (p i ) Technology 1: U 1 ( ,x 1,x 2 ) =  q 1 +(x 1 +α 1 β x 2 ) – p 1 Technology 2: U 2 ( ,x 1,x 2 ) =  q 2 +(βx 2 +α 2 x 1 ) – p 2

6. S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Example: IPv4 (Tech.1) IPv6 (Tech. 2) 6 Technology 1: U 1 ( ,x 1,x 2 ) =  q 1 +(x 1 +α 1 β x 2 ) – p 1 Technology 2: U 2 ( ,x 1,x 2 ) =  q 2 +(βx 2 +α 2 x 1 ) – p 2 –Cost (recurrent) of each technology ( p i ) –Linear Externalities (Metcalfe’s law) Normalized to 1 for Tech. 1 Scaled by β for Tech. 2 (possibly different from Tech. 1) α i, 0  α i  1, i = 1,2, captures gateways’ performance –Intrinsic technology quality ( q i ) Tech. 2 better than tech. 1 ( q 2 >q 1 ) –User sensitivity to technology quality (  ) Private information for each user, but known distribution

7. S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Low-def. video (Tech.1) High-def video (Tech. 2) 7 Technology 1: U 1 ( ,x 1,x 2 ) =  q 1 +(x 1 +α 1 β x 2 ) – p 1 Technology 2: U 2 ( ,x 1,x 2 ) =  q 2 +(βx 2 +α 2 x 1 ) – p 2 User sensitivity to technology quality (  ) –Private information for each user, but known distribution Low-def & High def video-conferencing service –Low-def has a lower price ( p 1 < p 2 ) but lower quality ( q 1 < q 2 ) –Video is an asymmetric technology HQ users get higher externality from each other ( β>1) Encoding is hard, decoding is easy –Low-def subscribers could display high-def signals but not generate them Converter characteristics –Simplex, asymmetric, unconstrained (α 1 β >1) –High/Low-def user can decode Low/High-def video signal –α i, 0  α i  1, i = 1,2, captures gateways’ performance

8. S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective User Adoption Process 8 Decision threshold associated with indifference points for each technology choice:  1 0 (x),  2 0 (x),  2 1 (x), where x=(x 1, x 2 ) U 1 ( , x) = 0 →  =  1 0 (x) - Indifferent to Tech. 1 U 2 ( , x) = 0 →  =  2 0 (x) - Indifferent to Tech. 2 U 2 ( , x) = U 1 ( , x) →  =  2 1 (x) - Indifferent between Tech. 2 and Tech. 1 Users rationally choose None if U 1 < 0, U 2 < 0 Technology 1 if U 1 > 0, U 1 > U 2 Technology 2 if U 2 > 0, U 1 < U 2 Decisions change as x evolves over time (myopic) x1x1 x2x2  ≥  1 0 (x) :U 1 ( , x) > 0 - Tech. 1 becomes attractive  ≥  2 0 (x) : U 2 ( , x) > 0 - Tech. 2 becomes attractive  ≥  2 1 (x) :U 2 ( , x) > U 1 ( , x) - Tech. 2 over Tech. 1

9. S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Diffusion Model 9 Assume a given level of technology penetration x(t)=(x 1 (t), x 2 (t)) at time t –H i (x(t)) is the number of users for whom it is rational to adopt technology i at time t (users can change their mind) Adoption dynamics: –Users differ in learning and reacting to adoption information –Diffusion process with constant rate γ< 1 –At equilibrium, H i (x*) = x i *, i  {1,2} –Determine H i (x(t)) f rom user utility function H 2 ( x(t))H 1 ( x(t)) H 1 ( x(t))= 1-  1 0 (x) H 2 ( x(t))= 0 H 1 ( x(t)) H 1 ( x(t))=  2 0 (x) -  1 0 (x) H 2 ( x(t))= 1-  2 1 (x)

10. S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Solution Methodology 10 P Q R1R1 R2R2 R3R3 R4R4 R5R5 R6R6 R7R7 R8R8 R9R9 x 1 =1 x 2 =1 0 0 x2x2 x1x1 Delineate each region in the (x 1,x 2 ) plane, where H i (x) has a different expression –There are 9 such regions, i.e., R 1,…, R 9 –Regions can intersect the feasibility region S Tech. 1 adoption level Tech. 2 adoption level S Each ‘region’ has different pairs of expressions of H 1 (x) and H 2 (x)

11. S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Computing Equilibria & Trajectories 11 Solve H i (x*) = x i *, i  {1,2} in each region Identify “candidate” equilibrium for each Region R k Candidates are valid only if they lie in their region Equilibria can be stable or unstable Trajectories: λ 1 and λ 2 can be positive, negative, or even complex

12. What are possible adoption outcomes? –Combinations of equilibria –Stable/ Unstable Adoption trajectories? –Monotonic vs. chaotic (cyclic) What is the role of gateways? –Do they help and how much? S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Key Questions 12

13. Theorem 1: There can be multiple stable equilibria (at most two) Coexistence of technologies is possible –even in absence of gateways Final outcome is hard to predict simply from observing the initial adoption trends S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Results (1): A Typical Outcome 13

14. Theorem 2: Gateways can help a technology alter market equilibrium from a scenario where it has been eliminated to one where it coexists with the other technology, or even succeeds in nearly eliminating it. Gateways need not be useful to entrant always! No gateways: Tech. 2 wipes out Tech.1 Perfect gateways: Tech. 1 nearly wipes out Tech. 2 S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Results (2): Gateways may help Incumbents 14 Tech. 1 adoption level Tech. 2 adoption level Tech. 1 adoption level Tech. 2 adoption level

15. Theorem 3: Incumbent can hurt its market penetration by introducing a gateway and/or improving its efficiency if entrant offers higher externality benefits (β>1) and users of incumbent are able to access these benefits (α 1 β>1) Theorem 4: Both technologies can hurt overall market penetration through better gateways. Entrant can have such an effect only when (α 1 β 1) Takeaway: Gateways can be harmful at times. They can lower market share for an individual technology or even both. S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Results (3): More Harmful Gateway Behaviors 15 Tech. 1 adoption level Tech. 2 adoption level Tech. 1 adoption level Tech. 2 adoption level

16. Theorem 5: Gateways can create “boom-and-bust” cycles in adoption process. This arises only when entrant exhibits higher externality benefits (β>1) than incumbent and the users of the incumbent are unconstrained in their ability to access these benefits (α 1 β>1) Corollary: This cannot happen without gateways, i.e., in the absence of gateways, technology adoption always converges Takeaway: Gateways can create perpetual cycles of adoption/ disadoption P.S: Behavioral Results were tested for robustness across wide range of modeling changes S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Results (4): More Harmful Gateway Behaviors 16 Tech. 1 adoption level Tech. 2 adoption level Tech. 1 adoption level

17. S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Limit Cycles: An Intuitive Explanation 17 Technology 1 Technology 2 Full-circle! α 1 β>1 Technology 1: U 1 ( ,x 1,x 2 ) =  q 1 +(x 1 +α 1 β x 2 ) – p 1 Technology 2: U 2 ( ,x 1,x 2 ) =  q 2 +(βx 2 +α 2 x 1 ) – p 2

18. Gateways can be useful to: –Promote coexistence & improve market penetration –Help lessen price sensitivity But, Gateways can be harmful too: –Hurt an individual technology –Lower Overall Market –Introduce Market Instabilities Analytical model is useful in: –Identifying scenarios for policy intervention –developing long-term strategic vision Qualitative results are robust to: –Different User Preference (θ) Non-uniform distribution (positively & negatively skewed β-distribution) Extended to externality benefits (i.e. θβx instead of θq+ βx) –Alternative Externality Models Non-linear externalities (Sublinear: x α, 0 1, Logarithmic: log(x+1 )) –Presence of Switching costs, Learning costs S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Conclusions 18

19. Models that do not consider individual user utility: Fourt & Woodlock (1960) – constant hazard rate model Bass (1969) - extension to include “word-of-mouth” effect Norton & Bass (1987) - successive generation of technology adoption Models that consider user utility function: Cabral (1990) – only single technology adoption Farrell & Saloner (1992) - homogeneous users Choi (1996) - extended Farrell & Saloner to include converters Joseph et al. (2007) – homogeneous users, model lacks dynamics What model do we really need? Competition between incumbent and entrant technologies Role of converters Heterogeneous users Adoption dynamics S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Related Literature 19

20. Network Infrastructure Choice: Shared Versus Dedicated Networks 1. Problem Formulation 2. Model & Solution Methodology 3. Key Findings & Examples 4. Conclusions S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Part 2: Outline 20

21. Emergence of new services require: –Network provider has to decide between: Common (shared) Network Infrastructure Separate (dedicated) Network Infrastructure Examples: –Facilities Management services & IT e.g. IT & HVAC systems –Video and Data services e.g. Internet & IPTV services –Cloud Computing e.g. Private (dedicated) cloud Vs Shared cloud –Broadband over Power lines Lack of Framework to evaluate choices: –Ad-hoc decisions (AT&T U-Verse versus Verizon FiOS) –Manufacturing Systems Literature: Plant-product allocation, optimal resource allocation S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Motivation 21

22. Plant-product allocation –How to allocate product demands to manufacturing plants –Effect of process flexibility in handling variable demand Jordan & Graves (1995) Graves & Tomlin (2003) E.K.Bish, Muriel, Biller (2005) Optimal Resource Allocation Fine & Fruend (1990) – firm’s optimal investment in flexible and dedicated resources J.A.Van Mieghem (1998) – role of price margins and cost-mix differential on flexibility benefits S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Related Literature 22

23. Two network services (technologies) –One existing (mature) service –One new service with demand uncertainty Sharing can create economies or diseconomies of scope in costs New service has demand uncertainty –Needs capacity provisioning before demand gets realized –Dynamic resource “reprovisioning” But some penalty will be incurred (portion of excess demand is lost) –Technology advances allow Reprovisioning (e.g., using virtualization) How critical is reprovisioning ability in choosing network design? –Compare networks based on profits S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Problem Formulation 23

24. S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Model Formulation 24 Basic Model: A Two-Service Model Service 1 (existing service) Service 2 (new service with uncertain demand) Three-stage sequential decision process Compare Infrastructure choices based on expected profits Reprovisioning Stage Capacity Allocation Stage Infrastructure Choice Stage Solve backwards

25. S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Model Variables 25 Provider’s profit depends on: Costs: Fixed costs Variable costs - grows with the number of subscribers (e.g. access equipment, billing) Capacity costs - incurred irrespective of how many users join (e.g. provisioning, operational) Cost ComponentService 1 Dedicated Service 2 Dedicated Shared Fixed Costsc d1 c d2 cscs Contribution Margin (grows with each unit of realized demand) p d1 p d2 p s1, p s2 Variable Costs (incurred irrespective of realized demand) a d1 a d2 a s1, a s2 Gross Profit Margin = p i - a i, i={s2, d2} Return on capacity = p i /a i

26. S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Solution (1): Reprovisioning Stage 26 Service 2 revenue: (i={s2, d2} for Shared and Dedicated respectively) i.when D 2 ≤ K i : R i (D 2 ≤ K i ) = p i D 2 – a i K i ii.when D 2 >K i : Reprovisioning Ability: A fraction “α” of the excess demand can be accommodated User contribution Capacity cost R i (D 2 > K i ) = (p i – a i )(K i + α(D 2 - K i )) A word about reprovisioning ability, α –Independent of the magnitude of excess demand –Captures feasibility of and latency in securing additional resources –So what do α =0 and α =1 mean?

27. S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Solution (2): Capacity Allocation Stage 27 Expected Revenue, E(R i |K i ), for a given provisioned level K i : Optimal Provisioning Capacity : For demand distribution ~U[0, D 2 max ]:

28. S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Solution (3): Infrastructure Choice Stage 28 Dedicated Networks: –Service 1 revenue: –Service 2 revenue under optimal provisioning: –Total profit: Shared Network: Infrastructure Choice: –Common if, else separate Profit from Service 2 Profit from Service 1

29. S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Choice of Infrastructure 29 Impact of system parameters: –Varying cost parameters affect the choice of infrastructure Shared to Dedicated (or Dedicated to Shared) Single threshold for switching n/w choice –Surprisingly, ad-hoc “reprovisioning” ability also impacts in even more interesting ways! Common is preferred over separate when Independent of provisioning decision Depends on provisioning decision Diff. in optimal capacity cost h(α)= Function of p i, a i, α, i={s2,d2}

30. S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Analyzing the effect of α on h(α) 30 Proposition 1: Increase in α benefits both shared and dedicated networks. (i) if increases in α benefits shared (dedicated) n/w more than dedicated (shared) (ii) if increases in α benefits shared (dedicated) more at low α and dedicated (shared) more at high α The value of h'(0) and h'(1) fully characterize the shape of h(α) Gross Profit Margin Return on Capacity

31. S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Results: Impact of Reprovisioning 31 GPM ded (p d2 -a d2 ) is sufficiently lower than GPM shr (p s2 -a s2 ) GPM ded > GPM shr i.e. (p d2 -a d2 ) >(p s2 -a s2 ) and ROC ded <ROC shr i.e. (p d2 /a d2 ) <(p s2 /a s2 ) GPM ded > GPM shr i.e. (p d2 -a d2 ) >(p s2 -a s2 ) and ROC ded >ROC shr i.e. (p d2 /a d2 ) >(p s2 /a s2 )

32. Developed a generic model that captures economies and diseconomies of scope between shared and dedicated networks Reprovisioning can affect the outcome in non-intuitive ways –Validates the need for models to incorporate this feature –Yields guidelines on how reprovisioning affects choice of architecture Identified key operational metrics to consider –Provides decision guideline Robustness: –Non-uniform demand distribution (positively & negatively skewed β-distribution) –Economies and diseconomies of scale –Different reprovisioning abilities for shared and dedicated networks ( α 1, α 2 ≠ α) S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Conclusions 32

33. Network Technology Adoption: (1)S. Sen, Y. Jin, R. Guerin and K. Hosanagar. Modeling the Dynamics of Network Technology Adoption and the Role of Converters. IEEE/ACM Transactions on Networking. 2010 (2)S. Sen, Y. Jin, R. Guerin and K. Hosanagar. Technical Report: Modeling the Dynamics of Network Technology Adoption and the Role of Converters. Technical Report. June, 2009. Available at http://repository.upenn.edu/ese papers/496/. (3)Y. Jin, S. Sen, R. Guerin, K. Hosanagar and Zhi-Li Zhang. Dynamics of competition between incumbent and emerging network technologies. In Proc. Of ACM NetEcon'08, pp.49-54, Seattle, WA, 2008. Network Infrastructure Choice: (4) S. Sen, R. Guerin and K. Hosanagar. Shared Versus Separate Networks - The Impact of Reprovisioning. In Proc. ACM ReArch'09, Rome, Italy, December 2009. (5)S. Sen, K. Yamauchi, R. Guerin and K. Hosanagar. The Impact of Reprovisioning on the Choice of Shared versus Dedicated Networks, Proc. of Ninth Workshop on E-business, WEB, St. Louis, MO, December 2010. S. Sen On the Adoption and Deployment of New Network Technologies: An Economic Perspective Bibliography 33 Thank You!