Summary

  • NANOG provides coordination infrastructure—role signals, appointment tools, a timed Peering Coordination Forum and tightly limited table materials—but no reviewed source shows that it negotiates, approves or guarantees a private interconnection.
  • The NANOG 89 forum lasted 90 minutes and accepted applications until 20 were received or the deadline; a successful applicant received a highboy table for up to two representatives, could display business cards and one white paper or one-sheet marketing page, received a rotating customised slide, and could offer no other giveaway.
  • NANOG's public, archived mailing-list rules draw an exact competition boundary around price fixing, boycotts, restrictions of trade, prices, discounts, sale terms, salaries, benefits, profits, margins, costs, market shares, territories, customer allocation, and customer or supplier selection or termination. Their published enforcement ladder is list-specific.
  • Privacy-safe accountability belongs at the venue NANOG controls: current eligibility and allocation rules, an event-wide competition boundary, coarse bands for applicants, capacity, introductions, follow-up and later self-reported activations, plus a statement that NANOG neither negotiates nor endorses nor guarantees a private agreement.

A highboy table with a narrow permission

The table was not a trading floor. NANOG described the Peering Coordination Forum on 16 October 2023 as a 90-minute session in which people working in peering could meet others attending NANOG 89. Applications were to remain open until 20 had been received or until 9 October, whichever came first. A successful peering representative was assigned a dedicated highboy table and could bring no more than two representatives.

The permitted materials were almost comically specific: business cards, plus one white paper or one-sheet marketing page. No other kind of giveaway was allowed. NANOG also promised a customised slide in a rotating display deck. The NANOG 89 agenda independently placed the forum at 17:00–18:30 and repeated the table and material arrangements.

These details define a genuine institutional surface. NANOG did more than leave a hotel corridor open and hope that useful encounters occurred. It allocated time, space, visibility and a small set of materials. Counterpart discovery became a recognisable part of the meeting. The format nevertheless stopped well short of publishing terms, matching traffic profiles, choosing counterparties or witnessing contracts.

Consider the encounter the rules permit. A representative notices a table, reads the rotating slide, recognises a potentially relevant location or capability, exchanges contact details and asks for a later conversation. The one-sheet may show enough technical or commercial overlap to justify follow-up. Nothing in that sequence sets a price, guarantees compatibility or authorises either person to bind an employer. The table can reduce search friction; it cannot complete the work.

That is NANOG's commercial boundary. A meeting for network operators would be less useful if business value were treated as contamination. The Internet is assembled from networks under separate control, so cooperation is part of operations. Commercial usefulness also makes the venue sensitive. Competitors can lawfully exchange much that enables interconnection, while coordination over prices, customers, suppliers or market division belongs on the other side of a bright line.

NANOG's public records show the two halves on different surfaces. The peering pages describe the physical permissions in unusually concrete terms. The mailing-list guidelines state detailed prohibitions around competitively sensitive coordination. The forum pages reviewed for this article do not visibly supply an equivalent event-wide statement connecting permission to meet with the subjects the institution does not facilitate. That public absence does not prove that staff give no private briefing or that no internal control exists. It means only that someone consulting the forum page cannot verify the complete boundary there.

Why competitors have to cooperate

Internet interconnection rests on a structural tension. Each network controls its own assets, customers, routing policy and commercial interests. It may compete with the network across the table for traffic, enterprise accounts, content delivery or geographic reach. Neither network can create global connectivity alone. Packets cross organisational borders because separately governed systems agree where and how to exchange traffic.

A 2015 study based on 38 interviews with engineers, peering coordinators and Internet exchange representatives described coordination as shaped by trust and distrust. Cooperation is necessary, but confidence is conditional. Interviewees assess technical capacity, reputation, expected traffic, operational responsiveness and the possibility that a counterparty's incentives may change. Social knowledge can help a coordinator judge whether the person across the table will answer during an incident as reliably as during an introductory conversation.

Large-scale agreement research supplies a different perspective. An OECD-hosted 2011 survey examined 142,210 carrier interconnection agreements and found informal arrangements pervasive. BEREC later hosted a 2016 update of the PCH survey. These studies do not show that NANOG caused a particular agreement, or even that one sampled agreement began at NANOG. They establish context: relationship infrastructure matters in a sector where consequential arrangements may remain informal.

Personal contact can therefore solve a real discovery problem. A coordinator with a dense professional network may already know who can answer, which meeting is worth scheduling and which technical issue is likely to halt a proposal. A newcomer may face generic addresses, slow replies and uncertain authority. A visible coordination forum can lower the first barrier by making relevant roles and availability easier to identify.

It cannot erase every asymmetry. An application process with a stated stopping point makes access to the visible surface finite, but the reviewed record does not establish that 20 tables were deployed. A two-person allowance shapes who can occupy an assigned table. Appointment tools may advantage people who already know what to ask and can describe a credible interconnection case. Repeat encounters may deepen trust. These are plausible access mechanisms, not findings of unfair allocation. They make current eligibility and selection rules worth publishing.

Technical and commercial questions also overlap. Do the networks share practical locations? Can both teams support the session? Are traffic characteristics acceptable? Who will handle escalation? Does the bilateral case require commercial terms? Some questions are technical, some commercial, and many cross the boundary. A trustworthy venue need not pretend the business side is absent. It should state what exchange it enables, what belongs in bilateral privacy and what kinds of competitor coordination are out of bounds.

Role recognition is not authority

Ashwin J. Mathew's 2014 UC Berkeley dissertation offers a close historical view of this coordination. Based on field observation at NANOG meetings and interviews with network personnel, it describes technical discussion, social relationships, business exchange and potential-customer contact occupying the same meeting. The research is qualitative, historical and partly anonymised. It cannot measure current table allocation or the rate at which introductions become interconnections. Its value lies in showing mechanisms that event pages tend to flatten.

One mechanism was a green badge marker identifying a peering role. Mathew describes it as a signal that could prompt a question about the network the person represented. In a crowded meeting, it helped another attendee recognise that a peering conversation might be relevant.

The marker did not say that its wearer could conclude a binding agreement. A role signal is not delegated authority, and the represented network remains responsible for its own approvals. The same is true of a table assignment, appointment or business card. Each identifies a possible route to a conversation; none proves authority to accept commercial terms, qualify a technical design or authorise a BGP session.

The fieldwork records both chance encounters and scheduled peering meetings. Those practitioners used PeeringDB information to evaluate possible counterparts. Interpersonal familiarity could help a coordinator bypass a slow organisational contact route and reach someone able to advance establishment work. That mechanism is credible: a known person can make a request legible inside a large organisation. The evidence does not show NANOG approving the connection or causing the eventual result.

One interviewee described NANOG as especially valuable for Tier 2 networks seeking peers and potentially reducing transit costs. This is an attributed view, not a measured savings result. It explains an incentive to attend: a new peering relationship may alter reliance on paid transit. It cannot establish how much money was saved, whether performance improved or whether the relationship would have occurred elsewhere.

Mathew also reports that sensitive company information and nondisclosure obligations made face-to-face exchange useful. The historic Peering BoF he studied was not recorded or broadcast, though it was open to NANOG attendees. Its format and norms cannot be assumed to match the current Peering Coordination Forum. Nor does the reviewed evidence establish whether current attendees still use the green role marker in the same way. The historical observation shows why role visibility can matter; present practice remains unknown.

The forum does not own the bargain

NANOG's historical self-description presents its meetings and membership as open to interested individuals and names operators, peering coordinators, vendors and researchers among attendees. The statement explains the community the institution seeks to convene. It does not prove that the room mirrors the North American network market, that all interested people have equal practical access or that attendees hold authority to represent operators collectively.

The Peering Coordination Forum turns that convening role into a particular service. It makes peering representatives visible, offers a fixed place, limits table materials and creates time for introductions. The NANOG 95 page again advertised the forum, described tools for one-to-one appointments and said applications remained open until 20 had been received or the deadline. That ceiling belongs to the reviewed NANOG 95 page. It must not be combined with another announcement that used a different number.

Historical notices show continuity without establishing a permanent rule. A February 2019 attendee message said tables for a future Peering Coordination event were available first come, first served. That was an allocation approach at the time, not evidence of today's rule. A February 2024 notice put the NANOG 90 forum at 17:00–18:30, confirming another 90-minute session. It did not report current selection criteria or outcomes.

Institutional responsibility should stop at the right boundary. NANOG controls the page, application, number of available places, table surface, published permissions and whatever event-wide instruction it gives. It can explain whether eligibility depends on role, whether excess demand is resolved in application order, through a relevance test, by lottery or through another stated method. It can publish available capacity and explain whether repeat participation affects priority.

Publishing a decision rule would not require publishing applicants. A short statement could distinguish minimum eligibility from allocation under excess demand, name any conflict-handling process and say when applicants will learn the result. If space is offered in application order, the page can say so for that event without turning a 2019 practice into a timeless rule. If staff use a relevance test, the criteria can be described without disclosing confidential business plans. If a lottery breaks ties, the unit entered in the draw should be clear.

These details matter because access to the organised surface is one of the few outcomes NANOG directly determines. They would also give an unsuccessful applicant a more useful explanation than silence while keeping names, proposed counterparties and strategy private.

The networks control what follows: whether to continue, which information to exchange, what technical tests to perform, whether locations and traffic are compatible, which legal form to use, whether payment is involved, how a route is configured and whether a relationship should later change or end. No reviewed source shows NANOG choosing counterparties, negotiating agreements, setting prices, approving BGP sessions, controlling routing policy or guaranteeing interconnection.

Those decisions unfold across distinct stages. Introduction identifies someone worth talking to. Bilateral negotiation tests whether commercial expectations can coexist. Technical qualification examines locations, capacity, traffic characteristics and operational practices. Organisational and legal approval establishes who may commit each network and on what terms. BGP activation implements an approved design. Routing choices and subsequent outcomes remain under each network's control. A table application or appointment cannot substitute for any later stage.

This separation also protects the meaning of a conversation that goes nowhere. Networks may decline because facilities do not overlap, expected traffic is unsuitable, staff capacity is limited or the commercial case is unpersuasive. None of those outcomes proves that the venue failed. Two representatives can also have a useful conversation without creating an interconnection. NANOG's task is to make the encounter trustworthy, not to pressure either side toward agreement.

The event's verbs should reflect that limit. Meet, discover, explore and schedule a follow-up describe what a coordination surface can support. Approve, deliver or secure a peering would imply control or assurance absent from the reviewed record. Precise verbs prevent convening power from being mistaken for authority over autonomous networks.

If access rules are hidden or unclear, the question belongs to NANOG because the institution owns the surface. If two networks cannot agree on terms, NANOG does not own their bargain. If a session later fails, the networks retain operational responsibility. A venue can be judged on trustworthy access without being credited or blamed for choices beyond its control.

The reviewed evidence also cannot show who gains most from the forum. Benefits might differ among large carriers, smaller networks, Internet exchanges, content networks, vendors, recruiters and other attendees. Historic observations of contacts and potential customers do not establish the present distribution. That unknown is another reason to report access and usage cautiously rather than treating open attendance as equal practical opportunity or representative authority.

An introduction is not an outcome

The temptation to over-credit a meeting is strong because the story is attractive: two engineers meet, recognise a match and turn a handshake into better connectivity. Sometimes an encounter may begin a valuable relationship. The reviewed evidence does not join a named NANOG introduction to a completed peering, a latency change, a reduction in transit spending, a new customer, a recruitment result or an increase in revenue.

A 2021 University of Central Florida thesis identifies NANOG networking events as one venue used by ISP administrators seeking new peering connections aligned with technical and business needs. It also argues that peer selection can take months and remain inefficient. That second finding keeps the first in proportion. An introduction may only begin qualification, internal approval, testing, legal work and implementation.

Networks have alternatives. They can use PeeringDB, email, Internet exchange meetings, private contacts, another operator forum or a direct commercial approach. A relationship that begins at NANOG might eventually have occurred through another channel. Even when an attendee says the meeting accelerated contact, separating NANOG's contribution from prior familiarity, technical fit and later work would require evidence not in the public record.

Process measures should therefore remain modest. NANOG could count how many applicants sought a table, how much capacity existed, how many applicants opted to report a new introduction, how many expected a follow-up and how many later placed their status in a broad activation band. These counts would describe use of the venue without claiming that the meeting caused a route, customer or saving.

Even activation requires care. Respondents might self-report that a new session became active within a coarse time band. NANOG should not name networks, list facilities, publish traffic volumes or infer financial value. A session can be established and later removed; an introduction can be useful even when the final answer is no; several conversations may contribute to one implementation. The measure would indicate follow-through, not return on investment.

The same restraint applies to customer and recruitment contact in the historical fieldwork. A meeting can place commercially relevant people together, but no reviewed count shows a customer win or hire caused by the forum. Credibility comes from reporting what the institution can know and labelling what it cannot.

The return reports the institution, not the private deal

NANOG, Inc.'s fiscal-year 2024 Form 990 offers a quantified view of meeting activity. The organisation identified itself as exempt under section 501(c)(3) and described its work as managing an educational and operational forum. Its filed mission language says the platform is intended to inspire, educate and empower a community working on the Internet. These are institutional statements on an information return, not independently measured results.

The filing reports three major conferences. NANOG 90 recorded 742 attendance instances: 629 in person and 113 virtual. NANOG 91 recorded 735: 673 in person and 62 virtual. NANOG 92 recorded 870: 808 in person and 62 virtual. The figures sum to 2,347 event-attendance instances, including 2,110 in-person and 237 virtual instances. Those totals are editorial calculations, not a count of unique people; one person could appear at more than one event.

The unit matters beyond a footnote. Three event totals can describe the volume of conference participation without revealing how many individuals attended once, twice or three times. They do not show how many people used a peering table, secured an appointment or even worked in interconnection. Nor can the in-person and virtual split be treated as an access measure for the forum, because the reviewed record does not join those attendance instances to its application or participation data. The filing gives a defensible scale for the organisation's conference activity.

It cannot supply the missing denominator for the narrower coordination surface.

The revenue classification is equally specific. The return reports REGISTRATION revenue of $1,244,354, with the full amount in related or exempt-function income. It reports MEMBERSHIP DUES of $64,526, also entirely in that column. The two figures sum to $1,308,880, exactly the reported program-service revenue.

On the expense side, the return reports $2,689,213 in total program-service expenses. Conference, convention and meeting expense was $2,147,253, of which $2,127,244 was placed in program services and $20,009 in management and general expense. The return reports total unrelated-business revenue of $0 and answers No to whether unrelated-business gross income reached $1,000.

These entries establish what NANOG filed. They do not show that the IRS examined and approved every meeting interaction or that private advantage was absent. The public reconstruction is derived from raw IRS filing data, but the underlying statements remain an organisation-filed information return rather than an IRS adjudication of each activity. Whether a tax authority reviewed or accepted every underlying interaction is unknown.

Form 990 instructions identify meeting or convention registration fees as a kind of program-service revenue. They distinguish related or exempt-function income as income from activities substantially related to exempt purposes. Separate IRS guidance explains that an income-producing activity must contribute importantly to those purposes; merely spending its proceeds on exempt work is not enough. IRS guidance generally defines unrelated business income through a trade or business that is regularly carried on and not substantially related to the exempt purpose, subject to exceptions.

Those principles explain reporting categories; they do not resolve every fact about NANOG. The organisation reported registration and membership income as related or exempt-function income and reported zero unrelated-business revenue. It would be wrong to translate those entries into proof that private advantage was absent, regulatory approval or a finding that each table conversation advanced an exempt purpose. Tax reporting classifies NANOG's income. It does not quantify the private value of an introduction, prospective customer, useful market intelligence or later peering.

Private value can coexist with public purpose

The tax figures invite a false binary. One view treats any private commercial benefit as evidence that an educational purpose has been corrupted. Another treats related-function reporting as proof that every meeting interaction is publicly beneficial. The evidence supports neither.

Network operations occur through organisations with private responsibilities. An engineer who learns a safer routing practice may improve an employer's service. A peering coordinator who meets a viable counterparty may reduce operational friction. A vendor may meet a prospective customer. Such benefits can coexist with technical education and community coordination; their existence alone does not establish corruption, capture or private inurement.

The useful question is whether the surface NANOG controls remains aligned with its stated function. Are access rules knowable? Does the forum separate introductions from negotiation? Are role signals clear without implying authority? Are competitively sensitive subjects bounded? Can people see that NANOG does not endorse a counterparty or guarantee a result? Can the institution account for use without collecting private terms?

Form 990 totals cannot answer those questions. Attendance instances reveal scale, not practical access. Registration revenue reveals resources, not the distribution of private gains. Meeting expense shows institutional cost, not whether a particular table was useful. Zero unrelated-business revenue states the filing position; it is not a public-benefit score.

Three ledgers should remain separate. NANOG's institutional ledger contains revenue, expense, attendance, application capacity and published rules. A private network's ledger contains staff time, travel, transit spending, customer strategy and implementation cost. The technical ledger contains routes, traffic, latency, resilience and incident response. A meeting may touch all three, but evidence in one cannot prove an outcome in another.

The distinction also prevents inflated credit. If one network later reduces transit costs, the change may depend on compatibility, bargaining, facilities, traffic, approvals and implementation long after the meeting. NANOG can fairly say that it created an opportunity to meet. On the reviewed record, it cannot claim the saving.

The public list draws a precise line

NANOG's mailing-list usage guidelines provide its clearest published competition boundary. The list is described as open, public, archived and community-moderated, intended for technical and operational exchange to an audience exceeding 10,000. Scale, permanence and public participation make this channel different from a private conversation at a peering table.

The rules prohibit private marketing initiatives and product marketing. They caution that product discussions are subject to antitrust law. More specifically, they prohibit posts encouraging or facilitating price fixing, boycott or illegal restriction of trade.

The prohibited coordination surface is detailed. The guidelines identify prices, discounts and terms or conditions of sale; salaries and employee benefits; profits, margins and cost data; market shares and sales territories; allocation of customers; and selection, rejection or termination of customers or suppliers as inappropriate subjects for agreement. This specificity matters. Do not discuss business would be unrealistic in an operator community. The published rule instead identifies categories in which competitor agreement or signalling can harm competition.

The enforcement sequence is equally concrete. It begins with a warning. After two warnings, a further violation can bring a 90-day restriction. A later violation can lead to permanent removal. The page also supplies a list-specific reporting route and moderation approach.

That ladder cannot be projected onto the Peering Coordination Forum, hallway conversations or private appointments. The source expressly governs the mailing list. Public, archived group discussion also differs from bilateral technical and commercial exploration. A lawful bilateral negotiation may require information that would be inappropriate to circulate among competitors as a group. An event statement should preserve that distinction rather than suggesting that every business subject is forbidden.

General FTC guidance explains the concern without turning it into an allegation. The agency says most trade-association activity is procompetitive or neutral, while exchange of current prices or identifiable competitively sensitive business information can raise antitrust concerns. It treats aggregated historical information managed by an independent third party as lower risk than identifiable current or future competitor information. This is general federal guidance. It neither labels NANOG a trade association nor finds a violation at its events.

Together, the list rule and general guidance point to a narrow event statement. People at the forum should be able to discuss technical compatibility, locations, operational requirements and whether a bilateral conversation should continue. NANOG should not facilitate competitor coordination over price fixing, boycotts, restriction of trade, prices or discounts, terms of sale, salaries or benefits, profits or margins, costs, market shares or territories, customer allocation, or customer and supplier selection or termination.

Why a private conversation can be safer

The strongest case against more disclosure begins with the purpose of the forum. A network cannot evaluate every possible interconnection in public. Traffic ratios may reveal customer mix. Locations can expose expansion plans. A planned facility or product may be confidential. Current transit terms can reveal negotiating position. Naming a prospective counterparty can distort a discussion before either side completes technical or legal review.

Historical fieldwork shows that nondisclosure duties and sensitive company information made face-to-face exchange valuable. Trust can allow a coordinator to say enough for the other side to evaluate a possibility without broadcasting it to competitors or customers. Repeat relationships can lower the risk that an exploratory question will be mistaken for a commitment.

Publishing every term could damage usefulness and competition at once. Transparency could become a signalling mechanism from which rivals infer pricing, capacity, territories or customer strategy. Smaller networks might be deterred from disclosing constraints to a possible partner. A recorded and searchable negotiation is not automatically more accountable.

Privacy is not a concession to wrongdoing. It can be an operating requirement. Networks need room to test compatibility, decline an offer and investigate sensitive issues without creating a permanent public narrative. NANOG should not collect details it cannot safely protect, and attendees should not have to disclose terms merely to prove that the event has value.

Accountability should therefore remain coarse. Report the capacity of the public coordination surface, not applicant identities. Report bands of introductions, not counterparties. Ask whether a follow-up was intended, not what price was discussed. If later activations are self-reported, combine them into broad time and count bands that resist re-identification. Do not publish traffic volumes, route details, customer names, contract form or commercial conditions.

Institutional claims require the same restraint. NANOG should not turn an aggregate activation band into a promise to future applicants, endorse one applicant as a superior peer or imply that table selection certifies technical or commercial quality. Confidentiality makes bilateral exploration possible; a clear disclaimer prevents institutional reputation from being mistaken for due diligence.

Publish the doorway, not the deal

The reviewed pages show a recurring coordination format but do not let the public reconstruct its present doorway. They do not visibly state current eligibility, how limited tables are allocated, whether repeat applicants receive different treatment, or whether selection is included in registration, separately paid, linked to sponsorship or governed by another condition. The 2019 first-come notice cannot answer those present-tense questions.

A current page could do so in a few sentences. It could state who may apply, how many spaces are available, what completes an application, how excess demand is resolved, whether repeat participation affects priority, and whether any fee or sponsorship relationship applies. None of that would expose a private discussion.

The page could include an event-wide competition boundary. It should welcome technical qualification and bilateral exploration while barring NANOG-facilitated competitor agreement on price fixing, boycott or restriction of trade; prices, discounts or sale terms; salaries or benefits; profits, margins or costs; market shares or territories; customer allocation; and customer or supplier selection or termination. Appropriate counsel should shape the wording for the event rather than copying a channel-specific rule mechanically.

That qualification matters. The mailing list is public, archived and moderated. A forum includes scheduled and chance bilateral encounters, some confidential. The event statement can preserve the substantive boundary without pretending that list moderation tools govern the room. It can state how concerns may be raised during the event and what role staff can perform.

The disclaimer should be just as direct: NANOG provides a coordination venue. It does not negotiate, endorse or guarantee a private agreement. It does not set prices, choose counterparties, qualify networks, approve BGP sessions or control routing. Participation and table assignment do not certify authority, compatibility or likely success.

The public pages reviewed here do not show all those elements. Their absence does not prove that there is no staff practice, private briefing or internal control. It shows only that an outside observer cannot verify the full boundary from the forum pages. Publishing the doorway would make NANOG's responsibility visible without exposing the deal.

Count access in bands, not bargains by name

An event-level or annual process note could support accountability without becoming a source of commercial intelligence. The first measures would concern demand and capacity: applicant bands, available tables, accepted places, withdrawals and no-shows. When counts are small, NANOG could combine adjacent events or suppress a result to reduce identification risk.

Immediate-process measures could remain optional. Respondents might report whether the forum created a new introduction, renewed a dormant contact or led to a scheduled technical follow-up. The form should not ask for the other network's name, intended locations, traffic estimates or terms. One respondent might report several introductions, so the publication should distinguish responses from people and avoid false precision.

A report would also need its response denominator. Ten self-reported follow-ups mean something different if ten respondents answered than if a hundred people were invited and most remained silent. NANOG could publish broad bands for invitations and responses, disclose that reporting was optional and decline to estimate results for non-respondents. Repeat respondents should not be silently counted as new people across events. These cautions do not make opt-in evidence useless; they prevent a convenient voluntary sample from becoming a success claim about every applicant.

Where the group is too small for a safe denominator, the institution can describe the question it asked and withhold the count.

A later check could offer coarse status choices: no follow-up, still evaluating, technical discussion, abandoned, or self-reported activation within a broad time band. Activation would need a neutral definition and an explicit statement that it was not independently verified. The measure should not ask why an effort stopped. Technical mismatch, commercial disagreement and strategic change can all be legitimate.

These counts should never become a league table. Ranking respondents by completed sessions would encourage selective reporting and disclose strategy. A success rate for a small pool might identify individual networks. Assigning financial value would invite unsupported assumptions about transit savings and revenue. The safest publication is a range with a clear denominator and privacy threshold.

Access data should not pretend that open application means equal opportunity. NANOG might report first-time and repeat applicants in broad bands, but it should not claim representativeness without a defensible population. Nor can the present record establish how benefits are distributed among different types of attendees. The aim is to examine recurring accessibility of the visible surface, not assign demographic or market meaning to an operational role.

The process note would remain useful even if few activations were reported. It might reveal excess demand, a high no-show rate, weak follow-up or a need for better appointment support. Those are features of the venue NANOG can improve. They do not require inspection of a contract or ownership of a route.

The business card is an invitation, not a mandate

The Peering Coordination Forum is best understood as arranged possibility. The highboy table says that a relevant representative will be available. The historic green role marker said that someone worked in peering. The business card provides a route to continue. The one-sheet offers enough information to decide whether another conversation may be worthwhile. None certifies a counterparty or creates a bargain.

That modesty fits interconnection. Competing networks must cooperate without surrendering autonomy. Trust may accelerate contact, while distrust and verification remain necessary. Informal arrangements may be common, but technical qualification, internal authority and operational implementation still belong to the networks.

NANOG's fiscal reporting shows an institution placing registration and membership income in related or exempt-function income while reporting substantial program-service and meeting expenses. It does not put a dollar value on introductions. The mailing-list guidelines show that NANOG can name prohibited competition subjects precisely. They do not establish the rules or enforcement method of each private exchange. Historical and independent research shows why face-to-face contact can matter. It does not prove a particular outcome.

The next improvement is neither radical transparency nor retreat from commercial exchange. It is to connect public statements at the surface NANOG owns: current access and allocation rules; a clear event-wide competition boundary; privacy-safe process bands; and an explicit no-negotiation, no-endorsement, no-guarantee statement.

Such a record would help newcomers understand the doorway, allow experienced coordinators to protect confidential discussion and prevent exaggerated credit or blame. It would also make the social boundary easier to see: people at the forum could distinguish useful bilateral exploration from a group exchange about prices, customers, suppliers or territories without treating all business conversation as suspect.

The institution's achievement is not the private bargain. It is the creation of a trustworthy place where autonomous networks can discover one another, decide whether a conversation is worth continuing and then leave the institution behind when negotiation begins. NANOG can own the integrity of that venue without owning the deal.

Sources