Innovative Capacity Allocations for All-IP Networks: A Network Economic Analysis of Evolution and Competition in the Internet Ecosystem (Freiburger Studien Zur Netzokonomie)

The Internet ecosystem is characterised by strong dynamics and rapid evolution. In view of the evolving range of content and applications that are to be delivered in an integrated fashion via general-purpose broadband networks, the requirements for traffic delivery and corresponding capacity allocation problems are becoming increasingly complex. The anticipated role of the Internet of things, 5G or virtual reality applications underscores how important it will be for broadband access service providers to flexibly meet these evolving demands in an economically efficient fashion. This book examines the evolution of, and competition within, the Internet ecosystem and analyses efficient capacity allocations in all-IP networks from a network economic perspective. In doing so, it critically appraises the role of network neutrality regulations. Cover Content Acknowledgements Papers List of Abbreviations 1 Introduction Part I The Internet Ecosystem: Evolution, Innovation, and Heterogeneity 2 The Internet from a Disaggregated Network Economic Perspective 2.1 The Economic Side of Layering: Disaggregating the Value Chain 2.2 The Technical Side of Layering: Modularity and the OSI and TCP/IP Reference Models 2.3 Reconciling the Economic and the Technical Sides of Layering: Towards a Layer Scheme for All-IP Networks 3 Internet Evolution: Towards Heterogeneity 3.1 From ARPANET to the Internet: A Historical Perspective 3.2 From Narrowband Access to the Broadband Internet 3.3 Broadband Evolution and the “Fiberization” of Access Networks 3.3.1 Broadband Network Evolution: A Dual Lineage 3.3.2 Upgrading the PSTN: From First-Generation xDSL to “Future-Proof” FTTH 3.3.3 Upgrading CATV Networks: From First-Generation Broadband to RFoG 3.3.4 Wireless Broadband Access: Towards 5G 3.3.4.1 (Cellular) Mobile Broadband: Fiberization and the Road Towards 5G 3.3.4.2 (Fixed) Wireless Broadband Access: WLAN, WiMAX, and Satellite Access 3.4 Evolution in the Broadband Access Ecosystem: A Brief Summary 4 Network Performance and Quality of Service 4.1 Quality of Service and Quality of Experience 4.2 Network Performance and Quality of Service 4.2.1 Why Location Matters: End-to-End Network Performance and the Role of the Geographical and Virtual Location 4.2.2 Delay 4.2.3 Jitter 4.2.4 Packet Loss Rate 4.2.5 Throughput, Bandwidth, and Data Rates 4.3 Heterogeneous QoS Requirements and the Evolution of Application Services 5 TCP/IP and the Best Effort Service Model 5.1 The End-to-End Arguments and the Best Effort Service Model 5.2 The Internet Protocol (IP) 5.2.1 The IP Packet Header: Beyond Addressing 5.2.2 IP Packet Forwarding and Routing 5.3 Some Fundamentals on Intra-Domain Routing 5.4 Some Fundamentals on Inter-Domain Routing: The Border Gateway Protocol (BGP) 5.5 An Introduction to the Transmission Control Protocol (TCP) 5.5.1 Flow Control 5.5.2 Congestion Control 5.5.3 TCP and Network Performance 5.6 An Alternative to TCP: The User Datagram Protocol (UDP) 5.7 Transport Layer Protocols: A Brief Summary 5.8 The Interconnection Ecosystem 5.8.1 Peering, Transit, and the Hierarchical Structure of the Traditional Interconnection Ecosystem 5.8.2 Innovations in Interconnection Agreements 6 Enhancing Performance in a Best Effort Internet 6.1 Over-Provisioning Strategies 6.2 Usage Constraints: Ad Hoc Quantity Rationing 6.3 Data Compression and Endpoint-Based Upper Layer Mechanisms 6.4 Overlay Networks and Location-Based Traffic Management 6.4.1 Overlay Networks and Content Delivery Networks 6.4.2 Edge Computing and Distributed Cloud Infrastructures 6.4.3 Internet Exchange Points, Remote Peering, and Growing Interconnection Diversity 6.5 The Limits of Enhancing Performance in a Best Effort Internet 7 Active Traffic Management and Enhanced Services on the Internet 7.1 Guaranteed Services and Flow-Based QoS: The Integrated Services Architecture and the Resource Reservation Protocol 7.2 Predictive Services and Class-Based QoS: The Differentiated Services Architecture 7.2.1 Per-Hop Behaviors (PHBs) 7.2.2 QoS Differentiations and the DiffServ Architecture: A Closer Look 7.3 Multiprotocol Label Switching, MPLS-TE and DS-TE 7.3.1 Some MPLS Basics 7.3.2 MPLS TE 7.3.3 DS-TE 7.4 The Generalized DiffServ Architecture Part II Congestion, Capacity Allocation, and Pricing in All-IP Networks The evolving demand for heterogeneous QoS levels is the key driver for the introduction of QoS differentiations. As described in Sections 6 and 7, a wide variety of technologically complementary traffic management mechanisms is available to facilitate the implementation of required service differentiations for the Internet. On the one hand, there are some mechanisms located at upper layers of the Internet protocol stack that can provide (scalable) QoS differentiations without requiring modification of the basic upstream architecture and infrastructure or interfering with best effort principles. Specifically, corresponding innovations like CDNs or endpoint-based application layer mechanisms enhance the capabilities of the basic TCP/IP-based Internet without requiring changes to the underlying architecture. In contrast, the introduction of QoS differentiations based on innovative traffic architectures such as DiffServ or IntServ/RSVP requires the adoption of new protocols. Therefore, corresponding changes of the underlying Internet architecture might give rise to significant coordination problems; both in terms of the protocol adoption at a technical level and with respect to the configuration of QoS classes and service class harmonization (see Section 13). Despite, or perhaps because of, their differences, different mechanisms for traffic management as well as related innovations take place in different layers. In fact, a variety of possibilities arise to combine their functionalities and capabilities. Even though substantial innovations in upper layer mechanisms seem to have gained increasing relevance in recent years, it was demonstrated in Section 6 that these mechanisms cannot provide full substitutes for active traffic management as provided by a more generalized traffic architecture. In fact, with the emergence of highly QoS-sensitive and mission-critical applications, for example those related to the Internet of Things or the Tactile Internet, the recognition of the importance of QoS-differentiated bandwidth capacities and mechanisms located at lower layers—potentially in conjunction with SDN/NFV-based network slicing mechanisms—seems to have reinvigorated the need for research on technical, economic, and regulatory frameworks capable of facilitating the digital transition and catering to the dynamically changing demands for QoS. 8 Scarcity, Network Externalities, and Two-Sided Markets 8.1 An Introduction to Scarcity and Economic Efficiency on the Internet 8.2 Network Externalities 8.2.1 Rivalry, Congestion, and Negative Externalities 8.2.2 Direct (Positive) Network Effects 8.2.3 Indirect Network Effects and Two-Sided Markets 9 Congestion and Pricing in the Internet 9.1 A Closer Look at Scarcity and Congestion on the Internet 9.1.1 Internet Congestion from an Economic Perspective 9.1.2 Algorithm-Based Capacity Rationing by the TCP 9.1.3 M/M/1 Queueing Models 9.1.4 Bandwidth Partitioning Approaches 9.1.5 Preliminary Conclusions on Congestion Modeling in Two-Sided Market Models 9.2 Internet Pricing and the Economics of Internet Congestion 9.2.1 A Primer on Usage-Independent versus Usage-Dependent Pricing 9.2.2 A Brief Survey of Traditional Internet Pricing Models 10 Economically Efficient Single-Channel Congestion Pricing 10.1 Some Basics on Optimal Congestion Pricing and Capacity Expansions 10.2 A Simple Model of Congestion Pricing for a Single-Class Internet 11 Towards Optimal Capacity Allocation in all-IP Networks 11.1 Paris Metro Pricing and its Implications for Multi-Channel Congestion Pricing 11.2 Pure Prioritization and Interclass Externality Pricing 11.2.1 Incentive Compatible Pricing and Optimal Investment: Social Welfare Maximization 11.2.2 Incentive Compatible Pricing and Optimal Investment: Profit Maximization 11.2.3 Interclass Externality Pricing 11.2.4 A Brief Assessment 12 Optimal Capacity Allocations in All-IP Networks 12.1 The Basic Idea 12.2 Capacity Reservation versus Packet Prioritization: On the Provision of Deterministic and Stochastic QoS Guarantees 12.3 Interclass Externalities and the Opportunity Costs of Different Types of Network Usage 12.4 Opportunity Costs and Interclass Externality Pricing for All-IP Networks 12.5 Optimal Capacity Allocations in All-IP Networks: Incentive Compatible Price and QoS Differentiations and Optimal Network Investment 12.6 A Brief Summary 12.7 An Example of a Price and QoS Differentiation Based on the Introduction of a Deterministic Traffic Class 12.8 Model Implications and Limitations: Extensions and Future Research 13 Implications for QoS-Based Interconnections and the Interconnection Ecosystem 13.1 Legacy Interconnections for an All-IP Ecosystem: A Growing Mismatch? 13.2 Challenges and Evolution in the Interconnection Ecosystem 13.3 A Summary and (Forward-Looking) Discussion Part III Ecosystem Evolution, Optimal Capacity Allocations and Network Neutrality Regulations 14 A Primer on Competition and Regulation of the Internet 14.1 The Many Facets of Competition in Network Industries 14.2 Remedying Market Power Problems: Regulatory Policy versus Active Competition Policy 15 A Sector-Specific Regulatory Framework for the Internet 15.1 Disaggregated Market Power Regulation for the Internet 15.2 Technical Regulations and Consumer Protection Measures 15.3 Universal Service Regulations 16 Network Neutrality Regulations and Optimal Capacity Allocations for the Internet 16.1 Network Neutrality: Normative Proposition and Regulatory Paradigm 16.2 On the Need for a Market-Driven Network Neutrality Concept 16.3 Network Neutrality Regulations in the U.S. and the EU 16.3.1 Network Neutrality Regulation in the U.S. 16.3.2 Network Neutrality Regulation in the EU 16.3.3 A Brief Summary of the Regulations in the U.S. and the EU 17 A Network Economic Analysis of Network Neutrality Regulations 17.1 Brief Overview of the Two-Sided Market Literature on Network Neutrality 17.2 A Critical Appraisal of Network Neutrality Regulations from A Network Economic Perspective 17.2.1 The Many Facets of Network Neutrality Regulations and Their Impact on Optimal Capacity Allocations and Innovation 17.2.2 Regulatory Market Splits and Structural Traffic Regulations 17.2.3 Reasonable Traffic Management and the Ban on Paid Prioritization 17.2.4 Minimum QoS Regulations and the Dirt Road Argument 17.3 Conclusions and Policy Implications 18 Summary References