Internet 3.0: Identifying Problems and Solutions to the ...
Internet 3.0: Identifying Problems and Solutions to the Network Neutrality Debate <ul><li>A Presentation at the </li></ul><ul><li>2007 Annual Conference of the </li></ul><ul><li>Association for Education in Journalism </li></ul><ul><li>and Mass Communications </li></ul><ul><li>Washington, D.C. </li></ul><ul><li>August 10, 2007 </li></ul><ul><li>Rob Frieden, Professor of Telecommunications Penn State University [email_address] </li></ul><ul><li>Web site : http://www.personal.psu.edu/faculty/r/m/rmf5/ </li></ul><ul><li>Blog site: http://telefrieden.blogspot.com/ </li></ul>
Goals of the Paper <ul><li>Identify the real access issues resulting from the Internet’s third major evolution. </li></ul><ul><li>Refute the claim that no problems have or will arise, but also that any type of discrimination will ruin the Internet. </li></ul><ul><li>Explain how “better than best efforts” packet routing can offer an worthwhile value proposition and enhancement of legacy routing arrangements. </li></ul><ul><li>But also identify scenarios where discriminatory and unlawful practices violate fair trade, antitrust and other laws. </li></ul><ul><li>Propose ISP reporting requirements as a transparent way to differentiate real vs. artificial network congestion. </li></ul>
Internet 3.0’s Challenges and Opportunities <ul><li>The Internet has evolved from a government incubated, specialized network into a commercialized “network of networks.” </li></ul><ul><li>To achieve greater market penetration, service diversification and greater commercial success the 3 rd generation Internet must have widespread broadband access. </li></ul><ul><li>Tier-1 Internet Service Providers (“ISPs”), largely owned by incumbent telephone companies, need to make sizeable facilities upgrades to support IPTV and other bandwidth intensive applications. </li></ul><ul><li>ISPs consider pricing and Quality of Service (“QOS”) discrimination/diversification as the best way to recoup infrastructure costs and generate greater profit. </li></ul><ul><li>Internet 3.0 will not have a single, “one size fits all” pricing and “best efforts” QOS. </li></ul><ul><li>The challenge lies in supporting diversification, including “better than best efforts” routing options, without adversely impacting the Web’s accessibility, competitiveness, synergies and serendipity. </li></ul>
Explaining the Concepts— Network Neutrality <ul><li>Advocates for network neutrality have identified threats, scenarios and some instances where an ISP has blocked, delayed, or otherwise thwarted the delivery of specific bitstreams. </li></ul><ul><li>Net neutrality advocates want to convert “aspirational” views of a democratic and nondiscriminatory Internet into enforceable rules that would restrict ISP flexibility in terms of pricing, service quality and offerings. </li></ul><ul><li>Advocates for net neutrality believe this principle should apply to all traffic routing arrangements, especially where content providers can secure delivery without direct payment to each participating carrier, e.g., Google pays for the an extensive and geographically diverse web presence, but benefits from the “peering agreements” its paid carriers have with other ISPs for content delivery to users of ISPs networks for which Google makes no direct payment. </li></ul><ul><li>Net neutrality advocates believe that the Internet has contributed to national productivity, economic opportunity and innovation in light of “best efforts,” end-to-end connectivity. </li></ul>
Explaining the Concepts— Network Flexibility <ul><li>Advocates for network flexibility reject constraints on their ability to price discriminate and recoup sizeable investment in broadband infrastructure. </li></ul><ul><li>They view net neutrality as thwarting competition and creating disincentives to invest in NGNs. </li></ul><ul><li>Advocates for net flexibility note that ISPs do not operate as common carriers and qualify for an “information services” deregulatory “safe harbor.” </li></ul><ul><li>As information service providers, ISPs have flexibly negotiated interconnection arrangements without evidence that any ISP or user group has faced concerted refusals to deal boycotts, or and other anticompetitive practices. </li></ul>
Telephony Cost Causation The caller usually triggers a complete end-to-end network setup using facilities provided by the originating carrier and other carriers secured by the originating carrier. Traffic measurement and tracking Metering and tracking likely. Parties Agree on a multilateral basis to divide cost and share toll revenues based on ITU Recommended model. <ul><li>Internet </li></ul><ul><li>Cost Causation </li></ul><ul><ul><li>Traffic types and routing vary making it difficult to use traffic flows for determining who should pay; conduit and content merge. </li></ul></ul><ul><li>Traffic measurement and tracking </li></ul><ul><ul><li>Possible, but does not necessarily indicate which party initiated the link and who benefits. Upstream and downstream flows often asymmetrical. </li></ul></ul><ul><li>Parties </li></ul><ul><ul><li>A connectionless protocol where many carriers may be involved in switching and routing packets on “best efforts” model; evolved from zero cost peering to a commercial hierarchy of peers and clients. </li></ul></ul>Telephony and Internet Models
User3 Carrier A Collects revenues Collects traffic Carrier B Receives half AR Terminates traffic User 1 User 2 User 1 User 2 User 3 Traditional Telephone Access Arrangement Carriers interconnect networks with the expectation of compensation for carrying and delivering traffic generated by another carrier.
ISP A Exchanges Traffic with other ISPs ISP B Collects revenues Requests and terminates traffic One-way (thick pipe) User 1 User 2 User 3 Large ISP A “peers” on a 0 cost basis with other large ISPs, but charges ISP B for access Small ISP B has to pay for access to ISP A’s network and peering access to other networks Two-way (thin pipe) Web 1 Web 1 Web 1 Internet Access
What Network Neutrality Advocates Want <ul><li>Network Neutrality advocates want to preserve the status quo where ISPs offer nondiscriminatory “best efforts” routing and reciprocal carriage agreements from other ISPs information service providers. </li></ul><ul><li>They like “All You Can Eat” service options, but do not fully recognize the different costs in serving “power users” (gamers, P2P networkers). </li></ul><ul><li>“ Nethead” philosophy about the Internet emphasizes lofty notions about ubiquitous access with less emphasis on cost recovery and analysis of cost causation. </li></ul><ul><li>Netheads favor zero payment Sender Keep All/Bill and Keep “peering.” </li></ul>
What Network Neutrality Opponents Want <ul><li>Because the major telecoms carriers own the major ISPs, a Bellhead management and telecoms cost recovery template predominates. </li></ul><ul><li>The telecom template has a route specific focus with comprehensive route tracking, usage metering and cost accounting. </li></ul><ul><li>Until recently Internet metering costs exceeded the benefits, and “tunneling” a complete end-to-end link was technologically difficult. Now both are doable, often accruing consumer benefits, e.g., March Madness real time streaming of tournament games to over 100,000 simultaneous viewers. </li></ul><ul><li>Network neutrality perceived as a constraint on price/QOS discrimination, harmful government oversight and a confiscation of private property </li></ul>
Calibrating ISP Rights and Responsibilities <ul><li>Justified Apprehension : Just as Enron employees found ways to create artificial congestion of electricity flows, ISPs have the capability and incentive to discriminate against a specific type of bitstream or generator of a bitstream without an operational justification. This includes blocking ports and “sniffing” bits to identify and block, delay or drop certain types of traffic, e.g., a competitor’s VoIP traffic. </li></ul><ul><li>ISPs can and should drop bits and deny service based on congestion and the inability to route bits. Net bias occurs when an ISP denies access even though ample capacity to switch and route the traffic exists. Net bias may not always violate contractual terms, because no contract may apply to participating upstream and downstream ISPs. </li></ul><ul><li>ISPs can and should offer end users different bandwidth and throughput speeds as well as different interconnection and access arrangements both upstream and downstream. </li></ul><ul><li>Net bias occurs when an ISP deliberately degrades service, drops packets and otherwise tries to punish a specific ISP or content source. </li></ul><ul><li>Unjustified Apprehension Net bias does not include services differentiation based on bandwidth, throughput, QOS, etc. </li></ul>
Net Bias Versus Reasonable Price and Service Discrimination <ul><li>Impermissible Net Bias </li></ul><ul><li>Deliberate Packet Loss </li></ul><ul><li>Creating (Enron-like) Artificial Congestion </li></ul><ul><li>Targeting Large Volume Content </li></ul><ul><li>Generators for Punishment or Extortion </li></ul><ul><li>Most Types of Port Blocking (but not to control spam and denial of service attacks) </li></ul><ul><li>Unilaterally Imposing Upstream and Downstream Rules That Violate Existing Service Level Agreements </li></ul><ul><li>Affiliate Favoritism That Violates service level agreements , Fair Trade and Antitrust Laws </li></ul><ul><li>Fees for Overriding Firewalls and Filters </li></ul><ul><li>Permissible Network Bias </li></ul><ul><li>Variable Bandwidth and Throughput </li></ul><ul><li>Bandwidth Partitioning </li></ul><ul><li>Metered Service </li></ul><ul><li>Better Than Best Efforts Routing </li></ul><ul><li>Akamai-type Enhanced Traffic Routing and Management </li></ul><ul><li>Special or Exclusive Content Deals </li></ul>
A Call for Best Practices <ul><li>Network flexibility in pricing, service provisioning and quality of service options can make economic sense. </li></ul><ul><li>However deliberate blocking or degrading traffic does not. ISPs should agree not to drop packets and create congestion when actual traffic conditions do not necessitate such action. </li></ul><ul><li>ISPs should be able to partition bandwidth and offer downstream end users and upstream ISPs different levels of bandwidth and QOS. However these options should be available to any user without retaliation and targeted degradation of service for refusing to pay for premium services. </li></ul><ul><li>Better than best efforts is not a contradiction, but existing interconnection and Service Level Agreements may restrict this option as might competition laws. </li></ul><ul><li>ISPs should fully disclose terms and conditions. Requiring transparency does not foreclose net flexibility. </li></ul><ul><li>The FCC should assert its ample “ancillary jurisdiction” to require ISPs to compile reports about network usage and congestion. </li></ul>