RIPE NCC and the economics of interconnection dependency

Summary

• Interconnection dependency is an economic condition, not a narrow routing topic: a network's ability to buy transit, peer, join exchange fabrics, enter cloud environments and retain customers depends on whether counterparties can trust its address evidence.
• RIPE NCC's most valuable role is not to decide traffic relationships. It is to maintain a thin, accurate ledger that protects uniqueness, continuity and verifiable authority for Internet number resources across its region.
• In a scarce-IPv4 environment, WHOIS/RDAP-style holder evidence, RPKI and ROA state, route-origin data, reverse DNS, abuse contacts and transfer history become bargaining materials at the peering and transit desk.
• Weak registry evidence acts as an interconnection tax. It appears as delayed turn-ups, stricter upstream filters, less favorable transit terms, cloud-admission reviews, reputation-cleanup costs and weaker customer portability.
• The burden is regressive. Large networks can absorb manual review, counsel and routing staff; smaller providers across the RIPE NCC service region may find that ambiguity converts scarce address space into dependence on stronger intermediaries.
• RIPE NCC should not become a traffic policeman, cloud gatekeeper or adjudicator of every commercial conflict. Its discipline should be narrower: keep the ledger coherent enough that private actors can make their own acceptance decisions without laundering disputes through registry discretion.
• The constructive standard is a ledger that makes the current holder, authorised origin, route evidence, RDNS delegation, abuse contact, transfer afterlife and correction path intelligible before interconnection negotiations begin.

The turn-up desk sees the ledger before it sees the traffic

The scene is ordinary enough to be missed. A regional network has ordered a cross-connect at an exchange facility. Its router is configured. The sales contract with a transit provider is nearly final. A customer migration is scheduled for the weekend. The autonomous system number is known, the prefix list has been circulated, and the engineer doing the turn-up has a narrow window in which to make the new path carry production traffic.

Then the questions begin. The upstream asks whether the customer may originate the prefix. The exchange route server wants route-origin evidence that matches the declared origin AS. A peer asks why an old routing entry still names a predecessor network. A cloud platform asks for a letter of authorisation and checks that the ROA is not going to mark the new announcement invalid. A mail deliverability vendor asks who controls reverse DNS. A security desk asks whether the public abuse contact reaches the operating team or a mailbox abandoned during a merger. The fibre is there. The router is there. The missing input is belief.

That is the first economic fact about interconnection dependency. A prefix is not commercially useful merely because it exists in a registry or can be announced in BGP. It becomes useful when independent counterparties accept the story attached to it: who is the recognised holder, who may originate it, who will answer complaints, who can delegate reverse DNS, whether the route-origin state is coherent, whether old reputation can be cleaned, and whether the planned change will survive automated filters and human review. The transaction is technical in execution but commercial in meaning. The route is a claim, and every counterparty decides how much evidence is enough.

RIPE NCC sits in the middle of this evidence economy. Its service region spans Europe, the Middle East and parts of Central Asia, covering more than 75 countries and over 20,000 local Internet registry organisations. Its public services include registry records, the RIPE Database, RPKI, reverse-DNS support and measurement or routing-information services such as RIPE Atlas and the Routing Information Service. Those facts do not, by themselves, define an institutional philosophy. They are exhibits. They show why the ledger is read by a wide class of carriers, exchange operators, platforms, customers, security teams, brokers and public-sector buyers that do not report to RIPE NCC yet still depend on the evidence it maintains.

The dependency has become sharper since IPv4 scarcity became normal. RIPE NCC's pool of freely available IPv4 addresses ran out in November 2019. Scarcity has made addresses more durable, more tradable, more likely to be leased, more likely to survive mergers and more likely to be embedded in customer contracts. At the same time, interconnection still works through many independent judgments. No single authority can force every upstream, peer, exchange fabric, cloud provider or enterprise customer to accept a route. Each network manages its own risk. Shared evidence is therefore valuable because it lowers the cost of many separate decisions.

The relevant lesson is not that RIPE NCC should decide who peers with whom. It should not. The lesson is that an accurate ledger reduces the price of being believed. If a small network can arrive at the turn-up desk with a current holder record, a matching ROA, coherent route data, working abuse contacts and RDNS authority, it negotiates over service. If it arrives with ambiguity, it negotiates over trust. That second negotiation is costly, slow and unequal.

Interconnection is a market for trust before it is a market for traffic

Interconnection is often described as a set of routing arrangements among autonomous systems: peering, transit, route servers, prefix filters, traffic engineering, settlement-free exchange, paid capacity and customer routes. That description is accurate, but it arrives late in the story. Before traffic moves, counterparties decide whether the network asking for acceptance is safe enough to believe. BGP announces reachability; it does not prove authority. The proof comes from records, history, contacts, route-origin validation, operating reputation and the commercial standing of the party making the claim.

This is why interconnection has always been partly a market for trust. A large incumbent can often substitute reputation for documentary neatness. Its name is known, its operating teams are findable, its customers are visible and its mistakes are likely to be treated as correctable. A smaller provider has less reputational surplus. It may serve a country, city, campus, industrial cluster or niche hoster with technical competence but little global recognition. For that network, the public record is not a ceremonial file. It is a portable credential that allows strangers to accept a routing relationship without turning every turn-up into a private investigation.

The same point holds for enterprise customers. A business that has invested for years in fixed firewall rules, supplier allowlists, remote-access policies, mail reputation, payment gateways and monitoring systems may want to move from one provider to another without renumbering. Its ability to do so depends on whether the new provider can carry the address range cleanly and whether the address evidence survives the move. If the registry record, ROA, route data, RDNS and contacts all align, the customer has a credible exit option. If they do not, the customer's address dependence becomes a switching cost that strengthens the incumbent.

Interconnection dependency therefore changes the bargaining surface. Transit terms are not only about price per megabit, port capacity and latency. They also reflect confidence in the customer's routing file. Peering decisions are not only about traffic ratios and geography. They reflect whether a route set looks stable enough to admit. IXP admission is not only about the payment of a port fee. It is also about whether route-server automation can handle the member's prefixes without creating risk for others. Cloud bring-your-own-IP programs are not only about platform preference. They require proof that the party importing addresses has authority and that the platform will not be blamed for carrying a dubious route.

The economic cost of weak evidence appears in mundane ways. A turn-up is delayed by a week. A peer accepts only a shorter prefix list. An upstream insists on provider-assigned space rather than customer-held addresses. A platform imposes manual review. A customer postpones migration because reverse DNS has not moved. A range with old abuse history requires a reputation-cleanup plan before a buyer will accept it. These costs are rarely booked as a registry expense. Yet they are caused, in part, by the quality of the public record and the ease with which counterparties can interpret it.

The decentralised internet can absorb many such frictions, but absorption is not the same as absence. If every network must prove routine facts separately to every counterparty, the market becomes biased toward large actors whose reputation or leverage can bypass review. A good ledger is the antidote. It turns private trust into common evidence. It does not eliminate business judgment. It makes business judgment cheaper.

The RIPE NCC record is bargaining evidence, not merely administration

A registry record compresses information. It tells a counterparty where to begin when assessing who holds a resource, who may speak for it, which network may originate it, how complaints should be routed and what continuity should be expected. The compression is valuable because interconnection is a high-volume environment. An upstream cannot conduct a corporate-history inquiry for every customer prefix. A route-server operator cannot manually examine every past allocation. A cloud platform cannot reconstruct every transfer chain. The market needs standardised evidence that is not perfect but is reliable enough to serve as the first pass.

In the RIPE NCC setting, that evidence has several layers. WHOIS/RDAP-style holder data helps counterparties identify the recognised resource holder and current contact surface. RPKI and ROAs help networks evaluate whether a particular origin AS has been authorised for a prefix. Routing-registry and route-origin data inform filters, route sets and operational expectations. Reverse-DNS delegation indicates whether the party claiming operational control can maintain name-resolution hygiene. Abuse contacts show where complaints should land. Measurement and routing-information services help observers see how routes behave over time. These layers are not all the same legal thing. They do not produce a single certificate of commercial truth. Their value lies in coherence across the file.

The word "coherence" matters. A counterparty may tolerate a missing field if the rest of the evidence is strong. It becomes much more cautious when different signals tell different stories. A holder name points to one entity while the letter of authorisation names another. A ROA authorises an old upstream while the new transit order declares a different origin. A routing entry is maintained by a consultant who no longer works for the holder. Reverse DNS still delegates through a previous provider. The abuse contact exists but does not answer. Each defect may have an innocent explanation. Together they raise the cost of acceptance.

Coherent registry evidence is especially important because private interconnection decisions often use automated systems. Route servers and upstream filters need inputs that can be turned into rules. Cloud platforms need repeatable admission checks. Security vendors need contact and reputation fields that can be processed at scale. Enterprises need documented assurance for change control. When public evidence is clean, automation helps smaller networks by reducing the need for personal familiarity. When evidence is noisy, automation can punish them because the machine cannot read the story behind the inconsistency.

That is why registry administration becomes bargaining evidence. The record does not determine every commercial outcome, but it shapes the starting position. A network with clean evidence asks for peering, transit or portability from a posture of normality. A network with weak evidence asks for an exception. Exceptions cost money because they require manual attention, risk acceptance and reputational cover. Someone must be paid, persuaded or indemnified to bridge the evidentiary gap.

RIPE NCC's public role should be judged against this market function. The ledger is useful when it protects uniqueness, accuracy and continuity without pretending to be a universal traffic court. It is harmful when it is too thin to support ordinary acceptance, or when it becomes so discretionary that routing and commercial disputes are pushed into registry judgment. The line is narrow, but it is the line on which much of the region's address economy depends.

IPv4 scarcity turns weak evidence into an interconnection tax

Scarcity changes the price of ambiguity. When addresses were easier to obtain, a network facing documentary friction could sometimes renumber, request different space or accept provider-assigned addresses while growing into its own later. That escape has narrowed. In the RIPE NCC region, IPv4 exhaustion means that scarce address resources are purchased, leased, inherited, reused, subdivided, financed and carried across corporate reorganisations. The address range often outlives the original business plan. Its evidence file must therefore outlive the original staff, provider, customer and system that first used it.

Weak evidence becomes an interconnection tax because every subsequent use must pay to clarify what should have been clear. A buyer of addresses must pay advisers to reconcile holder history. A lessee must persuade an upstream that the use is authorised. A provider onboarding a customer must align ROAs and route data before announcing the range. A company leaving a cloud platform must prove that the addresses can be carried elsewhere. A network with old abuse history must show that the responsible customer has changed and that complaints now reach the right team. None of these costs is the price of raw bandwidth. They are the price of doubt.

The tax is not levied uniformly. Large networks pay it with specialised staff, established contacts and the ability to keep redundant transit while records are corrected. Smaller networks pay it with delay, poorer terms and dependence. A small hoster in a secondary market may have no dedicated registry specialist. A university network may depend on historical allocations whose public records reflect old institutional structures. A regional provider may have one engineer who handles BGP, customer tickets, abuse complaints and procurement. When evidence is weak, these actors face the same proof demands as a multinational platform but with fewer resources.

The tax also appears in customer relationships. A business customer may have selected a network partly because it promised address portability or stable routing. If a later move is delayed by evidence gaps, the customer's operational risk rises. Mail reputation, payment processors, remote access, supplier allowlists and security policies are all tied to addresses. The customer learns that its dependency is not merely on bandwidth but on the quality of the address record. A network that cannot demonstrate control may retain the customer by default, not by merit. That is inefficient lock-in.

Scarcity magnifies reputation effects. A range used for abusive hosting, misconfigured mail or compromised services may carry risk long after the underlying customer changes. Cleaning that reputation requires operational work, but it also requires credible public evidence that the responsible party is reachable and the resource is under new control. If the record remains stale, counterparties may price the address range as tainted even after the network has corrected its operations. The tax is therefore dynamic: old ambiguity follows the resource into new deals.

The policy instinct in such conditions can be to ask the registry to police more. That instinct is understandable and dangerous. Scarcity does not turn a registry clerk into a landlord. It does not justify converting every commercial quarrel into a registry enforcement case. The better response is more disciplined evidence: make control, contactability, origin authority and correction history easier to verify, while leaving commercial acceptance to networks that must carry the risk.

RPKI and ROAs make authority machine-readable, but not complete

RPKI is one of the clearest examples of how registry evidence changes interconnection economics. A route-origin authorisation lets a resource holder state which autonomous system is authorised to originate a prefix. Validators can use that information to classify announcements. The result is not a complete theory of legitimacy, but it is a powerful reduction in a particular kind of doubt. A route that would otherwise require manual interpretation can be evaluated by machine-readable evidence linked to the registry system.

That reduction has commercial consequences. An upstream may require matching ROAs before accepting a customer route. A peer may reject RPKI-invalid announcements. An exchange route server may apply validation rules that limit a member's ability to propagate a prefix. A cloud provider may ask a customer to align ROAs before importing addresses. An enterprise customer may ask whether its provider's route-origin state will survive a planned migration. These are not abstract security preferences. They affect turn-up timing, supplier choice, redundancy and customer confidence.

ROAs also change bargaining because they make some mistakes more visible. If a network changes transit providers without updating origin authorisation, a legitimate commercial move can appear invalid. If a holder creates an overly narrow or overly broad authorisation, counterparties may worry about route leakage or unintended acceptance. If a lessee, managed provider or customer claims the right to originate a range but cannot align the ROA, the market may infer that authority is unclear. The inference may be wrong, yet it must be answered before the route is accepted.

The right way to value RPKI is therefore neither worship nor dismissal. It is not a certificate of full commercial entitlement. A valid route origin does not prove that every contract is sound, that customers consented, that reputation is clean or that the operator is prudent. Nor is it a cosmetic technical add-on. In a world of automated route filtering, RPKI is part of the evidence stack that determines whether a route can enter the market on ordinary terms. It makes one claim legible: this origin AS is authorised for this prefix by the party able to create the authorisation.

For RIPE NCC, the discipline is to support route-origin proof without converting the proof into traffic rulemaking. The registry should provide stable services, clear holder control, resilient validation pathways and practical correction routes. It should not decide that a private peer must accept a valid route, or that an upstream must reject every route whose commercial story is messy. Network operators own their risk. The registry's contribution is to make the underlying evidence less ambiguous.

This boundary matters because route-origin disputes can be tempting vehicles for mandate expansion. A conflict over a lease, takeover, customer move or routing accident may be presented as a problem that the registry should solve by disabling, revoking or favouring a route claim. Sometimes the registry must act within its narrow remit when records are false or authority is compromised. But it should avoid being drawn into deciding traffic relationships under the language of security. RPKI is strongest when it is a proof service, not a commercial veto.

Route data, filters and IXPs turn coherence into market access

Internet exchanges look neutral from a distance: a fabric, a port, a route server, a member list and a set of policies. In practice, they are evidence processors. An exchange must decide how routes are accepted, filtered and distributed. Route servers need input data. Members expect protection against leaks, hijacks and accidental propagation. The exchange cannot manually litigate every prefix announced across the fabric. It relies on structured signals and clear member responsibility.

That makes registry-adjacent route data part of market access. A member whose route set aligns with holder records, ROAs and operational contacts is easier to onboard. Its prefixes can be placed into filters with less manual effort. Its peers have fewer reasons to create exceptions. A member whose route data is stale or inconsistent may still be legitimate, but it becomes expensive to admit. The exchange operator must decide whether to spend staff time on review, ask the member to repair records, or suppress routes until evidence improves.

Upstream providers face a similar problem at larger scale. A transit carrier takes customer routes and carries them into the wider internet. If it accepts bad routes, it risks leaks, hijacks, reputational harm and customer complaints. If it is too conservative, it loses business. Registry evidence helps the carrier balance those risks. Good evidence lets the upstream say yes quickly. Weak evidence shifts the customer into manual review, stricter prefix limits, narrower accepted route sets or demands for provider-assigned space.

The economics are subtle because route filtering is usually presented as safety, not market structure. Safety is real. But safety systems also determine who can enter cheaply. A filtering regime based on coherent public evidence favours any network that can keep its records clean, including small providers. A filtering regime that requires private familiarity favours incumbents and large brands. The public ledger is thus a pro-competition input when it lets smaller networks meet the same acceptance standard without relying on a powerful sponsor.

This is particularly important across the RIPE NCC service region. Some markets have dense exchange fabrics and many upstream choices. Others have fewer interconnection points and higher dependence on regional carriers. A network in Amsterdam, Frankfurt, London or Paris may be able to route around one cautious counterparty. A network in a smaller market may not. The same evidence defect therefore has different economic consequences depending on geography. A delayed route update in a dense market is an inconvenience. In a thin market it may be a constraint on customer service.

Filters also create path dependence. Once a prefix is accepted through one upstream but not through others, the network may become dependent on that upstream's tolerance. The upstream can charge more, bundle services or impose terms because the customer has fewer credible exit paths. Clean public evidence widens the exit path. It gives the customer a file that can be carried to another carrier, exchange or platform. Weak evidence narrows it.

The registry cannot and should not dictate IXP or upstream filtering policy. The diversity of filtering choices is part of the internet's resilience. Yet RIPE NCC can lower the common cost by making the evidence inputs accurate, current and easy to reconcile. That is the modest, powerful role of a thin ledger.

Reverse DNS and abuse contacts are operational credit signals

Reverse DNS and abuse contacts are sometimes treated as low-status administrative details. They are not glamorous. They do not carry the intellectual prestige of routing security or the financial drama of IPv4 transfers. Yet they matter deeply in interconnection because they show whether a network can be reached, corrected and held operationally accountable after the route is accepted.

Reverse DNS is a practical signal of control. Mail systems, security tools, customers and peers use RDNS records when evaluating whether an address range is being operated coherently. A migration in which the prefix moves but RDNS remains with an old provider looks unfinished. A cloud import in which the platform carries traffic but the customer cannot align name evidence looks risky. A security review in which RDNS tells an old story can slow acceptance even when routing is technically correct. The issue is not that RDNS alone proves authority. It is that inconsistent RDNS invites doubt about who is actually operating the address range.

Abuse contacts perform a related function. They are not merely channels for complaint. They are signals that a responsible party can be found when things go wrong. Scanning, spam, credential stuffing, open resolvers, malware callbacks and route leaks all require escalation. A large network with imperfect public contacts may still be reachable through private channels. A smaller network cannot assume that privilege. Its abuse contact is part of its public credit profile.

This profile matters in commercial negotiation. A transit carrier may worry that a customer with poor contactability will generate complaints that fall back on the carrier. A peer may worry that operational incidents will be hard to resolve. A cloud platform may require a clear abuse path before allowing imported addresses. An enterprise customer may ask who will receive complaints that affect its own reputation. If the registry record provides a clean answer, the network reduces the risk premium attached to its routes.

Abuse contactability also affects reputation cleanup. Address ranges accumulate history. They appear in blocklists, geolocation databases, mail scoring systems, threat-intelligence feeds and private customer controls. Some history is deserved; some is obsolete; some is mistaken. To clean it, a network must be able to show that the current operator is reachable and that complaints will be handled. A stale contact makes cleanup harder because third parties cannot distinguish between a reformed address range and a neglected one.

Here again the registry's role is not to adjudicate every abuse complaint. RIPE NCC should not become the police desk for every spam run, phishing case or content dispute. It should ensure that the record supports contactability, correction and continuity. The difference is crucial. If the registry tries to decide which customers deserve interconnection after every complaint, it becomes a traffic gatekeeper. If it maintains reliable abuse and RDNS evidence, it allows counterparties to manage their own relationships on a common factual basis.

The low-status details are often where trust is most visible. A ROA may show that an origin is authorised; an abuse contact and RDNS delegation show that someone can live with the consequences. Interconnection is a continuing relationship, not a single route announcement. The mundane proof that someone is home is part of what makes the route acceptable.

Transfers have an afterlife at the peering desk

An IPv4 transfer does not end when the registry record changes. It has an afterlife in every interconnection relationship that must accept the new state. The buyer or recipient may need to update ROAs, routing entries, reverse DNS, abuse contacts, geolocation records, customer documentation, monitoring systems, blocklist appeals and cloud-admission files. If those changes are not sequenced well, the address range can carry a split identity: legally updated in one place, operationally old in another.

That split identity matters because counterparties rarely see the whole transfer file. A peer sees a new origin. An upstream sees a customer prefix that used to be announced elsewhere. A mail vendor sees changed RDNS. A security analyst sees old abuse reports. A cloud platform sees a customer asking to import space that still has historical fingerprints. Each reviewer forms a judgment from the signals available to them. If the signals disagree, the transfer is not commercially finished.

Post-closing ambiguity can be expensive. The buyer may have paid for addresses expecting immediate use, but the upstream refuses to accept routes until ROAs and route entries align. A customer may have contracted for continuity, but reputation systems still associate the range with the previous operator. A seller may remain reachable for complaints because abuse contacts were not updated. A platform may delay import because the holder evidence does not match the customer's name. The transfer has settled in the registry yet not in the market.

This is why transfer quality should be assessed by afterlife, not only by approval. A successful transfer is one whose evidence can be understood by the next counterparty. The registry record, route-origin state, contact surface and RDNS delegation should tell a coherent story that private actors can use without needing confidential deal documents. Some private detail will always remain private. But the public file should be clear enough that routine interconnection does not reopen the transaction.

The afterlife problem is distinct from the legal and financial mechanics of deal closing. The central issue is not escrow, taxation or legal opinions. It is whether the new holder can use the resource on acceptable terms after the deal. A perfectly documented closing can still leave the buyer with weak bargaining power if the operational evidence is stale. Conversely, a modest transaction with disciplined record maintenance can integrate smoothly because counterparties can see what changed and who is now responsible.

Reputation cleanup is part of the same afterlife. An address range may carry poor mail reputation, geolocation errors, fraud associations or security flags from earlier use. Registry evidence cannot erase history, but it can support a credible cleanup narrative. It can show that control has changed, contacts are current, RDNS is delegated correctly, and route-origin authority matches the new operating plan. Without that evidence, cleanup becomes a private pleading exercise, and platforms or peers may discount the range long after the underlying problem has ended.

RIPE NCC's contribution is again narrow. It need not certify that every transferred range is clean. It should make the current resource state clear enough for others to make their own judgments. That modest goal has large economic value because it reduces the deadweight loss between registry settlement and interconnection acceptance.

Cloud admission and enterprise portability depend on address evidence

Cloud platforms have made interconnection dependency more visible to enterprise buyers. A business may want to bring its own IP addresses into a cloud environment, move between cloud providers, connect through direct interconnect services, or keep customer-facing identifiers stable while shifting workloads. These decisions are sold as architecture choices, but they depend on evidence. The platform needs to know that the customer has authority to use the addresses and that announcing them through the platform will not create routing or reputation risk.

Bring-your-own-IP programs typically require holder evidence, authorisation documents, ROA alignment, route-origin coordination and operational contacts. A platform does not want to carry a prefix whose authority is disputed or whose route state will fail validation. Nor does it want to absorb complaints for addresses whose abuse path points elsewhere. The platform's admission decision is therefore a form of private interconnection review. It may rely on registry records even though it is not acting as a registry.

This affects bargaining. A customer with clean address evidence can treat cloud providers as substitutable to a greater degree. It can move workloads, keep customer endpoints stable and negotiate with credible exit options. A customer whose evidence is weak becomes more dependent on the platform or provider already willing to carry the range. The dependency may be invisible inside the architecture diagram, but it appears in procurement. The buyer cannot credibly threaten to move if every move requires uncertain manual review.

Enterprise portability also reaches beyond cloud. Banks, payment firms, logistics companies, SaaS providers, public portals, research networks and industrial systems may have address-based controls built into their partners' systems. Renumbering can be costly because counterparties must update allowlists, VPNs, monitoring, fraud systems and compliance controls. Provider-independent or portable addresses reduce that cost only if the new network can announce and maintain them credibly. Portability is therefore not a property of the prefix alone; it is a property of the prefix plus accepted evidence.

The risk is that weak registry evidence converts cloud and enterprise portability into a privilege for large actors. A multinational with counsel, networking staff and known platform contacts can push through review. A regional firm or public institution may not. If the public ledger is clean, the smaller actor can carry a standard proof file. If it is not, the smaller actor must borrow credibility from an intermediary. That is the interconnection tax in another form.

Cloud platforms are not villains in this story. They carry real risk when accepting customer addresses. They must protect their own networks and other customers. The better criticism is institutional: when public evidence is poor, platforms become de facto arbiters of address credibility because they control admission to a valuable environment. A thin and accurate registry ledger reduces that private gatekeeping by giving the platform and the customer a shared baseline.

The same logic applies to large enterprise customers choosing among connectivity providers. A provider that can demonstrate holder authority, ROA coherence, RDNS control and responsive contacts gives the customer more confidence that future moves will be possible. A provider that cannot do so may still offer good connectivity, but it offers weaker optionality. In an address-scarce world, optionality has value.

Small networks bear the heaviest fixed cost of doubt

The RIPE NCC region is large and uneven. It includes dense Western European markets with abundant interconnection options, sophisticated cloud and content presence, many exchange fabrics and mature network operators. It also includes smaller economies, frontier connectivity markets, specialised enterprise networks, research institutions, public-service providers, hosting firms and access networks whose staff and bargaining power are limited. The same registry ambiguity does not fall equally across that landscape.

Evidence costs are often fixed. A ROA must be corrected whether the network has one customer or one million. An old contact must be repaired whether the prefix supports a national carrier or a small city ISP. A route-filter review must be answered whether the operator is a global platform or a local hoster. Fixed costs are regressive: they consume a larger share of the smaller network's attention and margin. That is why registry quality is not only a technical public good. It is a market-access condition for smaller actors.

Consider a small provider serving a regional business community. It may need IPv4 addresses for web hosting, mail, remote access, VPNs, industrial systems and customer-facing services that still cannot rely on IPv6 alone. It may buy transit from one or two carriers, connect to a nearby exchange and sell continuity to customers who cannot tolerate address churn. If its public evidence is weak, each change becomes a negotiation. The provider must convince the upstream, the route server, the customer, the reputation vendor and perhaps the cloud platform. The cost of proving routine authority can be larger than the cost of the port.

Large networks have buffers. They have direct peering teams, registry specialists, legal counsel, automated record hygiene, account managers at platforms and a reputation that encourages counterparties to call before rejecting. Smaller networks often encounter the market through forms, ticket queues and route-server automation. Their evidence must travel without personal explanation. The public ledger is therefore more important, not less, for the actors with the least power.

Regional asymmetry compounds the problem. A network in a dense market can often try another peer, another upstream or another exchange location if one counterparty is slow. A network in a less dense market may face limited choice. Its dependence on the few available upstreams gives those upstreams more leverage. If weak evidence narrows choice further, registry ambiguity becomes a local competition issue. It may decide whether a smaller provider can offer credible services against an incumbent.

There is also a language and institutional-capacity dimension. The RIPE NCC region covers many legal systems, languages and corporate forms. Counterparties reading registry evidence may not understand local naming conventions, public-sector structures or historical reorganisations. A clear public record can bridge those differences. A vague record makes the weaker party explain its institutional context to every reviewer. That explanation is expensive and sometimes unsuccessful, even when the underlying claim is legitimate.

The fairness point should not be overstated. Networks that want interconnection must maintain their evidence. They cannot outsource every duty to the registry. But the registry can design its services so that routine correctness is easier, errors are visible, corrections are auditable and the current state is intelligible to strangers. Doing so reduces a fixed cost that otherwise falls hardest on the small.

The ledger must stay thin because traffic policing would corrupt it

The strongest argument for good registry evidence is also the strongest argument against registry overreach. Because registry evidence affects interconnection, it is tempting to ask the registry to solve more interconnection problems directly. A peer refuses a route; ask the registry to intervene. An upstream dislikes a customer; frame the dispute as resource misuse. A cloud platform demands stronger proof; ask the registry to bless the customer. A routing conflict emerges after a lease; ask the registry to choose which commercial story deserves traffic. Each request may sound practical. Together they would convert a coordination ledger into a traffic policeman.

That conversion would be a mistake. RIPE NCC's legitimacy rests on a narrow function: maintaining accurate, continuous and unique records for Internet number resources and supporting associated evidence services. It is not the buyer of transit, the operator of every exchange, the judge of every lease, the reviewer of every cloud import, or the commercial referee for every customer migration. If it starts deciding traffic relationships, it risks turning registry discretion into market power.

The danger is mandate laundering. A private actor may dislike a route for commercial, competitive, reputational or operational reasons. Instead of making that decision openly within its own network policy, it may try to push the conflict into registry action. The language will be public-spirited: security, accuracy, abuse, scarcity, trust. Some concerns will be real. But if every routing conflict can be recast as a registry matter, the ledger becomes a route-control venue. That would weaken the decentralised accountability of interconnection decisions.

A thin ledger is not a weak ledger. Thinness means precision of function. It means the registry records who is recognised, what resource is assigned or allocated, which contacts are responsible, which route-origin statements are authorised through the relevant system, which reverse-DNS delegation is in place, and how corrections can be made. It does not mean the registry ignores false data, compromised authority or broken contactability. It means it fixes evidence within its remit rather than turning evidence into traffic policy.

This boundary protects both networks and the registry. Networks keep responsibility for their own filtering, peering, transit and risk decisions. They can choose stricter or looser standards according to their business. The registry avoids becoming the venue for every commercial grievance. Resource holders receive a clearer bargain: keep the public evidence accurate, and you can ask counterparties to judge you on that evidence; but do not expect the ledger to force others to carry your traffic.

Scarcity makes the boundary harder because addresses have economic value and disputes become sharper. The more valuable the resource, the greater the incentive to use registry procedures as leverage. That is precisely why the boundary matters. Scarcity does not make the registry a landlord with broad discretion over every valuable use. It makes the registry's narrow duty more important: protect the shared evidence layer so the market can transact without central traffic control.

The best institutional posture is therefore disciplined modesty. RIPE NCC should be demanding about the accuracy and continuity of its ledger, but restrained about substituting its judgment for private interconnection decisions. The goal is not to make every route acceptable. It is to make the facts around that route clear enough that acceptance or refusal belongs where it should: with the networks carrying the risk.

A constructive standard for interconnection evidence

If weak evidence is an interconnection tax, the constructive question is what evidence should be cheap to assemble before a network reaches the turn-up desk. The answer is not a grand new bureaucracy. It is a disciplined, serviceable file that counterparties can read quickly and that resource holders can maintain without hiring an army of specialists.

The first element is current holder authority. A counterparty should be able to see the recognised resource holder, the responsible organisation and the relationship between the holder and the party asking for interconnection. Where a managed provider, lessee, customer or platform is involved, the public record will not reveal every private term, but it should not tell a contradictory story. The holder's authority should be legible enough for a letter of authorisation or route request to make sense.

The second element is route-origin coherence. ROAs should match the intended origin AS. Routing-registry data and route sets should not point to obsolete arrangements. Changes should be timed so that a legitimate migration does not become RPKI-invalid or fail a filter because one evidence layer moved before another. The market does not need perfection; it needs planned sequencing and visible responsibility.

The third element is operational contactability. Abuse contacts, network-operation contacts and administrative contacts should reach responsible teams. The standard should not require exposure of unnecessary personal data. It should require that complaints, emergencies, routing issues and verification requests do not disappear into dead mailboxes. Contactability is part of address credibility.

The fourth element is reverse-DNS continuity. RDNS delegation should move with operational control and customer expectations. It should not remain stranded with a prior provider after a migration or transfer. Because RDNS affects mail, security tooling and customer trust, it should be treated as part of the interconnection file rather than an afterthought.

The fifth element is transfer afterlife. When address resources move, the public evidence should support the next use, not merely record that a registry change occurred. Holders should have practical checklists for ROAs, route entries, RDNS, abuse contacts, customer notifications and reputation cleanup. RIPE NCC does not need to certify each downstream correction; it can make the expected evidence chain clearer.

The sixth element is auditability of correction. When records are wrong, the path to correction should be understandable, documented and proportionate. Counterparties do not need to see private documents, but they benefit from knowing that the current record resulted from a real correction path rather than arbitrary change. Auditability lowers suspicion without exposing sensitive material.

The seventh element is clarity about boundaries. Every public service should make clear what it proves and what it does not prove. A ROA proves route-origin authorisation in a specific sense; it does not prove all commercial rights. An abuse contact proves a channel; it does not prove good conduct. A holder record proves registry recognition; it does not settle every private claim. Boundary clarity protects the ledger from overuse and helps counterparties combine evidence sensibly.

Such a standard would help the whole region. It would allow peers, upstreams, exchanges and platforms to automate more safely. It would help small networks prepare interconnection files without guessing what each counterparty will ask. It would help buyers and customers distinguish real risk from stale paperwork. It would let RIPE NCC improve the ledger's market value without taking control of traffic relationships.

The economics of acceptable terms

The phrase "acceptable terms" is important. Interconnection dependency is not only about being accepted or rejected. It is about the terms on which acceptance occurs. A network with weak evidence may still obtain transit, but perhaps only from a carrier willing to charge more for review and risk. It may still peer, but perhaps with fewer networks or only through bilateral arrangements that take longer. It may still use cloud services, but only after manual admission. It may still retain customers, but with weaker promises of portability. Acceptance on bad terms is still dependency.

Economists would describe the difference as a bargaining problem shaped by information costs. When both sides can verify basic facts cheaply, negotiation focuses on price, capacity, service level, geography and commercial fit. When basic facts are costly to verify, the stronger side asks for compensation: delay, indemnity, restrictive terms, higher prices or control over the customer's addresses. The weaker side may accept because it needs connectivity more urgently than the counterparty needs the deal.

The address ledger reduces information cost by making core facts public and standard. It does not eliminate asymmetric information. The resource holder still knows more about its contracts, customers and operating competence than the counterparty does. But the ledger narrows the uncertainty. It tells the counterparty that certain baseline claims are not merely self-assertions. In a decentralised market, that narrowing can be enough to improve terms.

The effect is clearest for customer portability. A customer that can show clean resource evidence has a credible threat to move. Credible exit changes how providers behave. It disciplines pricing, improves service quality and reduces the hold-up risk that appears once a customer's systems depend on stable addresses. A customer whose address evidence is messy may still own or control valuable resources, but it cannot easily turn that control into bargaining power. Its exit threat is discounted.

The same effect appears for smaller networks seeking upstream diversity. A provider with clean evidence can solicit offers from multiple carriers. A provider whose routes require special explanation may receive fewer bids. The difference between two serious upstream offers and one reluctant offer can determine margin, resilience and investment. The ledger therefore shapes not just technical acceptance but the competitive structure of downstream markets.

There is a broader regional implication. RIPE NCC's region includes markets where network competition, data-centre development, cloud adoption and enterprise digital services depend on the ability of smaller and mid-sized actors to interconnect credibly. If registry ambiguity favours incumbents and large platforms, the cost is not only borne by the holder of a single prefix. It appears in higher prices, weaker resilience and slower local service development. A clean ledger is not industrial policy, but it is infrastructure for competition.

The proper measure of success is therefore avoided friction. How often can a legitimate holder or operator present its address evidence and move through interconnection review without bespoke explanation? How often do transfers settle into usable routing without a second round of doubt? How often can a customer take addresses to a new provider without discovering that the evidence file was never portable? These questions are more useful than counting forms completed. They measure whether the ledger is doing economic work.

The tax of doubt should be visible, not hidden in operations

One reason interconnection dependency is under-analysed is that its costs are hidden inside operations. A delayed peer is logged as a provisioning issue. A cloud-admission review is treated as platform bureaucracy. A rejected route appears as a filter problem. A stale RDNS delegation becomes a mail-deliverability ticket. An abuse contact failure becomes a security complaint. A transfer afterlife problem becomes a customer-migration delay. Each event has its own local cause, but the common pattern is weak shared evidence.

Making the tax visible would improve governance. RIPE NCC need not publish commercial judgments or private disputes to learn from friction. It can examine where holders struggle to maintain evidence, where contacts fail, where RPKI changes create migration risk, where RDNS delegation lags, where transfer recipients need better aftercare, and where route data is hardest to reconcile. The point is not to shame networks. It is to understand where the ledger is ceasing to lower transaction costs.

Resource holders also have responsibilities. They should treat registry evidence as part of interconnection readiness, not as paperwork postponed until a provider asks. Before a turn-up, migration, transfer, cloud import or customer portability event, they should know whether the holder record, ROAs, route data, contacts and RDNS tell the same story. The cost of cleaning records after a failed review is higher than the cost of maintaining them in advance.

Counterparties should be precise about what they need. An upstream that rejects a route should identify whether the concern is RPKI state, route data, holder evidence, contactability, prefix length, reputation or local policy. A platform that asks for proof should distinguish registry evidence from contractual assurance. An exchange that filters routes should make its evidence inputs clear. Precision prevents every refusal from becoming a vague suspicion and helps holders correct the right layer.

RIPE NCC's own standard should be neither maximalist nor passive. It should not promise that a clean record guarantees interconnection. It cannot. Networks remain free to reject routes for their own reasons. But it should recognise that record quality affects market access and that poor evidence imposes costs beyond the registry office. The ledger should be designed for the practical readers who use it: peers, transit carriers, exchange operators, cloud platforms, customers, security teams, buyers and lenders.

The final institutional point is continuity. Running networks matter. Address changes, transfers, security upgrades and record corrections should be managed so legitimate services do not become collateral damage. Continuity is not an excuse for bad data. It is the reason to fix data carefully. A registry that protects continuity while improving evidence lowers the cost of trust. A registry that either ignores ambiguity or overreacts with broad discretion increases it.

RIPE NCC's interconnection importance therefore lies in what it should not become as much as in what it should do. It should not become a landlord of scarce IPv4, a traffic policeman, a platform-admission board or a substitute court for every routing quarrel. It should remain a thin, accurate ledger whose public facts are good enough for the market to use. In a scarce-address world, that modest role is not small. It is the difference between a prefix that can bargain on acceptable terms and a prefix that must buy credibility from someone stronger.