Who Pays What to Whom in the Internet Ecosystem?

Internet interconnections and compensation arrangements have been based on voluntary terms and conditions after government underwriting stopped. In a commercial context, funds flow on the basis of traffic volume, but also bargaining power (individually, or collectively held).

Last mile ISPs have significant bargaining power in light of their control of access to end users ("eyeballs" to advertisers and content providers alike).  While we could debate about the robustness of competition and last mile broadband options (functional equivalency), few would disagree that most consumers select one and only one last mile wireline provider (where available) for traffic requirements exceeding 10 Gigabytes.  The FCC termed last mile ISPs as "terminating monopolists," appropriate insofar as most consumers are willing to pay for and rely on a single wireline last mile ISP.  Consumers might also subscriber to a wireless broadband carrier, but data caps force conservation and create incentive to retain a wireline option offering unlimited data, or very high caps like Comcast's 1 terabyte monthly cap, plus Wi-Fi tethering.

In the context of platform/double-sided markets model, last mile ISPs have 2 compensation/revenue streams available: 1) last mile broadband subscriptions and 2) payments from upstream CDN and content providers.  While I can see how a credit card company might need to offer free cards, or even pay car users with airline miles and rebates, last mile ISPs have not given up imposing tiered service rates, primarily based on transmission speed (and not data volume like wireless carriers).  Last mile ISPs now want to increase the compensation received from upstream players.

Netflix, Google and other large volume generators of content have been free riders only if one considers content providers/distributors as solely responsible for compensating downstream carriers handling more traffic than generating upstream.  But Netflix is not the sole revenue source: last mile ISPs used to rely primarily (if not exclusively) on their retail subscribers' monthly payments.  The last mile ISPs want to maximize revenues on BOTH sides of their platform and I don't see the same financial constraints like that incurred by credit card companies.

Netflix blinked first, had payer's remorse, but the commercial negotiation process generates winners and losers.  For my part, I don't see how commercially negotiated compensation arrangements trigger network neutrality concerns, unless and until the FCC has jurisdiction to apply Title II, which it now has, plus evidence that the arrangement is discriminatory and/or unreasonable as defined by the FCC.  Bear in mind that the FCC has eschewed requiring tariffs and doesn't apply the prohibition of paid prioritization upstream from the last mile ISP.

So in large part, it seems to me that a maturing Internet ecosystem has diversified from the traditional peering/transiting dichotomy into a variety of hybrid arrangements.  Has such diversification harmed competition and/or consumers?  I'm not sure that it has, even when zero rating has the potential to influence consumers' content choices by injecting a possible cost avoidance factor.

Latest Publications: Conflict in the Network of Networks: How Internet Service Providers Have Shifted From Partners to Adversaries

The Hastings Communications and Entertainment Law Journal (Volume 38, No. 1 pp. 63-90) has just published an article of mine that examines the peering process.  I show that peering "partners" no longer have a common and shared mission.  Instead the parties become opponents on compensation and interconnection arrangements as occurred between Netflix and Comcast.

The article reports on the FCC's concerns about peering and the prospect for government oversight established in the 2015 Open Internet Order.  I suggest that the FCC refrain from intervening unless and until a stakeholder files a formal complaint based on the assertion that the parties cannot achieve a a commercially negotiated arrangement.

Cable Retransmission/Channel Placement Negotiations and Commercially Reasonable Internet Connections

            Back at the drawing board, Chairman Wheeler and staff have attempted to find the sweet spot where ISPs can negotiate paid traffic prioritization so long as it’s “commercially reasonable.”  Libertarians and a lot of other observers would conclude that all commercial negotiations reach a reasonable outcome between two willing parties.  So absent coercion or evidence of an unfair—okay call it unreasonable—trade practice, the negotiation should produce a mutually beneficial outcome.

            Such outcomes do not prevent one side from exercising superior bargaining leverage.

            In broadcaster-cable television retransmission consent negotiations, the former enjoys a superior bargaining position for two reasons: 1) broadcasters have exclusive access to “must see” television such as the regular season of professional football and 2) cable operators face severe restrictions on their ability to negotiate with a distant broadcaster if the local station imposes unreasonable demands. So arguably the deck is stacked in favor of broadcasters.

            What does the Commission do in this situation?  Nothing for two reasons: 1) the Commission lacks specific statutory authority to impose terms and conditions; and 2) the Commission wisely refrains from interfering with “marketplace driven” negotiations knowing that eventually the parties will reach closure, particularly after the regular NFL season begins.  The Commission limits its intervention to defining what constitutes good faith negotiations.

            I acknowledge that the consequences of regulatory reticence to act can more significantly harm consumers when ISPs cannot come to terms.  The pain threshold arrives almost immediately when access to the Internet cloud becomes congested, or when specific sites become inaccessible.  Many would assert that reliable and neutral Internet access has more significance than whether cable television subscribers can watch a football game. 

            Similarly the D.C. Circuit Court of Appeals has instructed the FCC that it lacks jurisdiction to supersede cable operators’ channel placement and content tiering decisions. Absent a “voluntary” commitment, as occurred when Comcast agreed to limits on its channel placement freedom, the FCC cannot mandate neutrality and fairness.  Comcast can place its owned and operated Golf Channel on the basic tier and relegate the Tennis Channel to a more expensive tier viewed by fewer subscribers.   Was this a commercially prudent decision, or one designed to disadvantage the Tennis Channel?  The court in effect said it does not matter.

            The FCC has a model in retransmission consent and case precedent that it may not consider applicable.

           

Post Network Neutrality Feud Number 1: The Netflix (Traffic) Jam

            As you know, the D.C. Circuit Court of Appeals has invalidated network neutrality requirements that impose common carrier requirements.  In this blog and elsewhere I predicted an uptick in disputes between content providers and distributors in the absence of unquestionable authority for the FCC to intervene if necessary. 

To be clear I favor commercial negotiations that typically resolve interconnection compensation disputes.  However, I also suggest that the FCC have authority to resolve intractable disputes as a referee and mediator.

So along comes another dispute between Netflix and retail ISPs such as Verizon and Comcast.  See Drew FitzGerald & tzGerald   BiograShalini Ramachandran, Netflix-Traffic Feud Leads to Video Slowdown, The Wall Street Journal (Feb. 19, 2014); available at: http://online.wsj.com/news/articles/SB10001424052702304899704579391223249896550?mod=WSJ_hp_LEFTTopStories.

This really should not come as a surprise, even as retail ISPs already receive compensation on both sides of their two-sided market: 1) 3 digit margin monthly broadband retail subscriptions; and 2) transit payments from ISPs, particularly Content Distribution Networks for Netflix such as Level 3.

Retail ISPs want a third revenue stream on some notion that content sources, such as Netflix, are “bandwidth hogs” who should be throttled, or alternatively hit up for direct payments.  In particular it must tick off senior management at ISPs, owned by cable television companies, to see Netflix offer a $7.99 value proposition when cable content bundles are 10-15 times as expensive.

I agree that a direct payment should flow from Netflix if and only if it directly interconnects with a retail ISP.  If Netflix were to stop using CDNs and seek to interconnect directly with ISPs providing the last mile delivery Netflix surely should pay including the significant electricity used to power onsite proxy servers. 

But are retail ISPs right to demand payment from both the directly interconnecting upstream ISP/CDN and even farther upstream from the content source?

I don’t think so, but there’s nothing stopping retail ISPs from trying.  Apparently Verizon and others can degrade Netflix traffic delivery—intentionally or not—without much consumer pushback.  When consumers don’t get high resolution Netflix content, they do not even know whom to blame.  Has Netflix done something wrong, or has the last mile carrier?  Who operates the weakest and inferior link when multiple ISPs participate in the complete end-to-end routing of traffic?

Until retail ISPs lose customers or the debate in the court of public opinion expect more interconnection compensation disputes to arise and possibly mess with your Internet access experience.

Thoughts on AT&T's "Toll Free" Data Delivery Service

            AT&T wireless has announced a campaign offering content providers the opportunity to pay for access to end users so that the downloading does not debit customers’ data plans.  Some think this constitutes double or triple dipping, because AT&T already receives compensation from end user broadband subscriptions and ISPs with which the last mile provider has a peering or transit agreement.

            What’s “fair” and “unfair” in this discussion highlights the complexity when a transaction combines content and conduit.  Efforts to separate the two and price them out can raise fairness concerns.

            Consider a transaction where the separation occurs more readily: paying for an item, e.g., book content from Amazon, and paying for shipping, e.g., postal delivery of a hard copy book, or download from Amazon to an e-reader.  Vendors often bundle “free shipping.”  For its part Amazon wants customers to use “free” wi-fi for the download, but the company does pay for wireless carrier delivery when necessary to cellular equipped e-readers.

            A content distributor, like Netflix, wants to avoid having to pay the U.S. Postal Service for physical delivery of DVDs and similarly the company wants to avoid having to pay last mile retail ISPs. For Netflix Internet delivery can save the company money, because end user retail Internet access subscriptions can include the book delivery within the “free delivery” monthly data allotment.  Bear in mind that until now Netflix consumers didn’t have to think about downloading costs thanks to unmetered data delivery.

            With monthly data caps, the delivery aspect becomes more visible and potentially costly to both the content provider and the consumer.  Cost recover gets murky, because even before companies like AT&T want to hit up Netflix for “toll free” downloading there already are two sources for offsetting carrier download costs.

            Offering Netflix the opportunity for not causing a downloading debit adds a potential third revenue stream. Is this an extension of the double-market economic construct like that envisioned by Chairman Wheeler?  Bear in mind that even now Internet compensation arrangements are negotiated between directly interconnecting parties that barter access or secure payment.

Tracking New Models and Conflicts in Web Interconnection and Delivery

            You might have an interest in my work to understand the diversification of web interconnection and content delivery models, largely driven by the substantial increase in streaming video and the proliferation of Content Delivery Networks.  ISPs have devised many new deviations from the traditional peering/transiting dichotomy including: use of Internet Exchange Points by Tier-2 ISPs, paid peering, CDN surcharges, equipment co-location, e.g., Netflix Open Connect Network; “specialized networks” and Intranets/ Multiprotocol Label Switching and non-carriers like Google securing Autonomous System identifiers.

            Some retail ISPs also want to increase to three the number of payers for last mile content delivery.  Currently end users pay monthly Internet access subscriptions and directly interconnecting, upstream carriers pay when traffic for delivery well exceeds what the retail ISP can or will hand off for upstream carriage.  The targeted third revenue source does not directly interconnect, but constitutes a major source of content, e.g., Netflix.

            I’m working on a paper that examine existing and likely future interconnection disputes with an eye toward identifying where conflicts will arise and whether commercial negotiations can reach closure on a timely basis.  Here’s a link to slide pack summarizing the paper:
http://www.personal.psu.edu/rmf5/New%20Models%20and%20Conflicts%20in%20the%20Interconnection%20and%20Delivery%20of%20Internet-mediated%20Content.ppt.

Mission Critical Bits and Pay to Play Net Bias

             The proliferation of video content options via the Internet raises questions about what ISPs can and should do to offer “better than best efforts” to enhance quality of service.  Is this an opportunity for “pay to play” extortion, or welcomed quality of service discrimination?  One might assert the lack of a need for service prioritization in light of the absence of network congestion, but as bandwidth intensive, video demand increases does this conclusion make sense?

            Video content often qualifies as “mission critical bits” whose delivery must arrive on time, or the streaming content freezes and evaporates.  For example, Netflix and its subscribers expect each and every link to work with sufficient switching, routing and transmission capacity to deliver packets on a timely basis.  Few Netflix subscribers would stick with the company if suddenly full motion video streams became slide shows of random frames. 

            Increasingly ISPs want to secure surcharge payments from companies like Netflix to guarantee timely packet delivery.  So on top of the double-sided market where ISPs already receive payments from end users and upstream carriers, such as Content Distribution Networks, a third revenue stream should flow further upstream from content providers like Netflix.  Is this being greedy, particularly in light of what ISPs markets to subscribers?  Bear in mind that traditionally both peering and transit agreements involved directly interconnecting carriers, not ones further upstream or downstream.

            Broadband end users expect their $50-75 monthly subscriptions to cover the cost of access without a surcharge to them and others for the privilege of accessing full motion video sites.  ISPs already have the option of charging more for high volume users.  ISPs: send “broadband hogs” fruit packets and a higher bill, not throttled service.  ISPs also tier service based on bit transmission speed.  Are they entitled to more compensation from the sources of content that motivate broadband subscriptions in the first place?

New Publication--The Mixed Blessing of a Deregulatory Endpoint for the Public Switched Telephone Network

        Telecommunications Policy soon will publish my paper entitled The Mixed Blessing of a Deregulatory Endpoint for the Public Switched Telephone Network

        Here's the abstract:

            Receiving authority from a National Regulatory Authority to dismantle the wireline public switched telephone network (“PSTN”) will deliver a mixture of financial benefits and costs to incumbent carriers.  Even if these carriers continue to provide basic telephone services via wireless facilities or the Internet, they will benefit from the likely substantial relaxation of common carriage duties, no longer having to serve as the carrier of last resort and having the opportunity to decide where and what services they will provide going forward.  On the other hand, incumbent carriers may have underestimated the substantial financial and marketplace advantages they also will lose in the deregulatory process.

            Incumbent carriers often obscure or dismiss as insignificant the substantial privileges and benefits accruing from their status as telecommunications service providers.  Common carrier responsibilities include duties to interconnect with other carriers, provide service on transparent and nondiscriminatory terms and offer some low margin services.  But this legal status also guarantees wireline local exchange carriers in many nations access to annual universal service funding, zero or low cost access to rights of way and radio spectrum, accelerated depreciation and other tax benefits, the ability to vertically integrate throughout the “food chain” of telecommunications services and dominant status in the administration of telephone numbers, standard setting and other policy issues.  Incumbents will strive to capture deregulatory benefits while retaining the many benefits previously reserved for common carriers.

        This paper will identify the potential problems resulting from the decision by the United States Federal Communications Commission (“FCC”) to grant authority for telecommunications service providers to discontinue PSTN services.  The paper also will consider whether in the absence of common carrier duties, carriers providing telephone services, including Voice over the Internet Protocol (“VoIP”), voluntarily will agree to interconnect their networks.  The paper will examine Internet peering and other types of network interconnection with an eye toward assessing whether a largely unregulated marketplace can ensure ubiquitous access to PSTN replacement services. 

        The paper concludes that private carrier interconnection models and information service regulatory oversight may not solve all disputes, or foreclose price discrimination for functionally the same type of service.   Recent Internet interconnection and television program carriage disputes involving major players such as Comcast, Level 3, Fox and Cablevision, point to the possibility of increasingly contentious negotiations that could result in balkanized telecommunications networks with reversed or reduced progress in achieving universal service goals.  The paper also concludes that rural access to VoIP and other voice communications services could end up costing significantly more than what urban residents pay, an efficient, but politically risky outcome.

Yale JOLT Article on Internet Access Regulation

     The Yale Journal on Law and Technology has recently published my article entitled Rationales for and Against Regulatory Involvement in Resolving Internet Interconnection Disputes (14 Yale J.L. & Tech 266 (2012); available at: http://yjolt.org/rationales-and-against-regulatory-involvement-resolving-internet-interconnection-disputes.
     Here's the abstract:

      This Article will examine the terms and conditions under which Internet Service Providers (“ISPs”) switch and route traffic for each of several links between a source of content and consumers. The Article concludes that the Federal Communications Commission (“FCC”) may lack direct statutory authority even to resolve disputes based on its determination that Internet access constitutes an unregulated information service.  Additionally the FCC may appropriately forebear from regulating, because sufficient competition favors industry self-regulation.

       Despite substantial reasons not to intervene, the FCC nevertheless might have to clarify its understanding of what subscribers of retail ISP services can expect to receive. Under truth in billing and other consumer safeguards the Commission might require ISPs to explain what an Internet subscription guarantees not only in terms of transmission speed and downloading capacity, but also what subscribers can expect their ISPs to do when receiving content requiring downstream   termination.

      The Article concludes that both customers of content services, such as Netflix, and retail ISP subscribers expect their service providers to guarantee delivery of movies and all sorts of Internet traffic respectively. For physical delivery of DVDs Netflix must pay the U.S. Postal Service and for delivery of streaming bits Netflix must pay one or more ISPs. But for Internet traffic involving two or more ISPs, the Article examines whether other retail ISPs providing last mile delivery of content violate their service commitments to subscribers by demanding additional payment from upstream carriers.

Intranets and the Cloud: The Lack of Functional Difference Between Comcast’s Xbox and Regular Broadband Traffic

             Those clever folks at Comcast want us to believe that there is a functional difference between routing Netflix traffic to an Xbox 360 terminal versus conventional Internet routing via a cable modem.  See http://blog.comcast.com/2012/03/xfinity-on-demand-on-xbox-and-your-xfinity-internet-service.html.  Comcast states that routing to the Xbox 360 “will not travel over the public Internet and is delivered in much the same way as we deliver your video service to your set top box.”  Besides the company notes that only the worst of the worst bandwidth hogs would ever exceed the company’s generous 250 Gigabyte download monthly quota.

             As a threshold matter the company wants to differentiate the Internet and the Internet cloud from individual networks.  But of course we understand cloud computing and the Internet as requiring seamless connectivity between and among disparate but interconnected networks.  If we focus on the first of possibly many links from Netflix downstream to subscribers’ Xbox 360 terminals, are we to infer that Comcast owns to leases every possible segment in the complete link from content source to content end users?  If the answer is yes, does that mean that Netflix’s primary Content Distribution Network, Level 3 never handles Xbox 360 terminal traffic?

            While Comcast has a robust and broadly distributed network I seriously doubt that the company exclusively uses its own network assets to secure the complete routing of Netflix traffic to Xbox users.  Nondisclosure agreements block any transparency in understanding how carriers route traffic, but I cannot believe Comcast would give up the opportunity to continue charging Level 3 a surcharge for Netflix traffic and would gladly free Level 3 and other carriers of the long haul carriage burden.  In other words I suspect Comcast receives Netflix traffic in the same way—via Level 3 mostly—for both Netflix traffic destined for a cable modem and end user computer and Netflix traffic destined for an Xbox 360.  If anyone know this to be incorrect, please set me straight.

            I assume that the only difference in routing occurs in the so called last mile routing from the point where Comcast receives Netflix traffic.  This last link hardly qualifies for the broadsweeping pronouncement by Comcast that Xbox360 traffic is the functional equivalent of cable television content.

No Free Lunch in Internet Peering or Transit

              Like many of you, I am keenly following the Comcast-Level 3 dispute and am trying to make sense of it all.   The dispute confirms several universal principles about Internet traffic routing that have passed the test of time:

1)         Consumers pay Internet Service Providers (“ISPs”) a monthly subscription with the expectation that the fee covers access to available content, i.e., the conduit.  As the World Wide Web evolves and content options diversify to include full motion video, consumers simply expect their ISPs to make sure the download distribution pipes are sufficiently robust to handle high bandwidth requirements and commensurately large monthly download volume.  Cable modem service agreements may have a cap on downloading per month, but consumers generally assume “All You Can Eat” access rights, plus the expectation that video streaming will work, i.e., no blurring, frozen frames, or blue screens.

2)         Because upstream requests for content are narrowband and because the typical consumer downloads much more content than he or she uploads, ISPs serving end users, such as Comcast, typically will have a large traffic imbalance with more downstream traffic to deliver than upstream traffic that the end user serving ISP might want other ISPs, such as Level 3, to handle.

3)         Until such time as Comcast’s “Television Anywhere” takes off and generates lots more traffic that Comcast will need other ISPs to handle—whether on a peering or transit basis—Level 3 vastly contributes to Comcast’s download “surplus” delivery burden to end users.  Of course Level 3 replaces another content distribution network so the total volume of Comcast’s downloading burden does not change in the short term. However, in the context of peering and transit between Comcast and Level 3, the traffic volume relationship changes with a greater imbalance resulting from the new Netflix traffic Level 3 now delivers to Comcast.

4)         The Comcast- Level 3 dispute distills to a disagreement over whether and how much either should pay in light of changed traffic patterns.  Because the parties already have traffic agreements, modification of terms might require additional payments from Level 3 to Comcast, absent Comcast’s need for Level 3’s upstream transmission services.  Of course Comcast does need the services of Tier 1 ISPs like Level 3, but until Comcast starts distributing lots more of its cable television video product over the Web, Netflix downloading to Comcast subscribers will predominate.

5)         Cooperative ISPs typically align inbound and outbound peering traffic with an eye toward creating a balance, but either or both ISPs might also want to expand transiting services as these paid arrangements are based on the unlikelihood of balanced traffic loads.   Digital Society Policy Director George Ou reports that Comcast and Level 3 have both peering and transit agreements; see http://www.digitalsociety.org/2010/12/video-level-3-versus-comcast-peering-dispute/.  George lays blame on Level 3 for expecting Comcast to absorb the newly increased volume of traffic delivered to it by Level 3 without additional payment by Level 3, or the offer of additional free upstream capacity.

            Reasonable people can disagree as to the mutual exclusivity or substitutability of peering versus transit.  George considers the two types of traffic arrangements mutually exclusive and has chided me for thinking that the parties could recalibrate both to mitigate the traffic imbalance if they wanted to.  See http://www.digitalsociety.org/2010/12/many-analysts-wrong-on-comcast-versus-level-3/. 

            The Comcast- Level 3 dispute confirms that there is no such thing as a free lunch.  It also highlights disagreement over who has to pay when consumers’ download requirements increase with full motion video access.  George considers it a nonstarter for Comcast to raise end users cable modem rates, despite a vast increase in the value proposition created by IPTV.   Some economists consider it a given that Comcast has the “right” to demand compensation from both sides of its market position, upstream from Level 3—and possibly the real instigators of greater bandwidth requirements Netflix and Google—and also downstream from end users, i.e., cable modem subscribers, co-conspirators with Netflix and Google.

            Bottom line: one or more players in the Internet “network of networks” will have to pay for greater capacity.  Early on in the Internet’s development, avoiding payment strategies were depicted as “hot potato routing.”  Carriers unwilling to upgrade facilities to accommodate greater demand sought to hand off traffic as soon as possible.  Level 3 has no such option of passing the packets off to several different carriers for the last mile to end users.  Comcast knows this and true to form the company exploits its position to the fullest extent possible.

Comcast’s Demand for a Video Surcharge From its Level 3 “Peer”

According to Level 3, a major long haul Internet Service Provider, Comcast has demanded a “recurring fee” when Level 3 hands off movie and other high capacity video traffic for delivery by Comcast to one of the cable company’s subscribers.  See http://lb.vg/46734.  This demand warrants scrutiny, perhaps less in the context of Network Neutrality and more in terms of further diversification (unraveling) of the peering process.

I will leave to others the advocacy for and against another Comcast innovation in non-neutrality.  The company must consider its merger with NBC a done deal as it continues to maintain a high profile for provocative actions that raise rates to rivals and subscribers alike.

My interest lies in the evolution of peering, a process that used to be symmetrical and largely uniform between similarly sized ISPs.  Under the old school model, Level 3 would have similar peering agreements with Comcast as with other national cable operators.  Likewise Level 3 would have symmetrical terms for the carriage of its traffic downstream via a “peering partner,” such as Comcast, and for Level 3’s carriage upstream of traffic originated or passed onto Level 3 by Comcast.  So under the old model, if Comcast wants to single out a particular type of traffic for a surcharge payment from Level 3, then all things being equal at least in terms of traffic volume, Level 3 could require a similar payment from Comcast. 

           

Under the traditional peering model, if traffic volumes are roughly equal, the surcharge Level 3 would have to credit for payment to Comcast would be offset by a roughly equivalent credit to Level 3 for video traffic originated over the Comcast network, or transiting through it. If Comcast unilaterally has demanded and received the right to a video delivery surcharge without a reciprocal payment to Level 3, then Comcast either has eliminated the conventional symmetry in peering, or much more traffic originates or transits through Level 3 networks destined for Comcast subscribers than Comcast hands off to Level 3.  The fact that Level 3 has capitulated to Comcast’s surcharge demand points to a significant imbalance in traffic flow and commensurate negotiating clout.

           

Much of the Network Neutrality debate has focused on end user access, while peering changes are negotiated agreements about access upstream from end users.  The peering process is obscured by Nondisclosure Agreements and the lack of readily available data on traffic flows.  Comcast may be engaging in a shakedown designed to handicap competitive alternatives to Video on Demand, but the possibility exists that the company is responding to unequal traffic volumes.  We may never know which.