A Registry Without a Balance Sheet: The Low-Value Premise of Early Coordination

Summary

• Early coordination made technical assignment state its administrative object: unique values, reserved capacity, delegated blocks, registered-network information, and a centrally updated aggregate record.
• NSF later procured registration as an operating service with named providers, schedules, budgets, performance terms, reports, incremental funding, and controlled treatment of fee income.
• RIPE NCC’s completed 1995 accounts exposed the cost and financing of regional administration, yet those institutional categories supplied no price for the number resources being administered.
• Scarcity entered IAB policy in 1990; Geoff Huston’s 1994 argument then separated queue priority, assignment duration, administrative cost, and resource price, opening accountability questions that technical records alone were not designed to answer.

The Administrative Object: Assignment State and Coordinated Uniqueness

The administrative object of early number coordination was a state in a shared technical namespace. RFC 790, issued in September 1981, compiled values assigned from several numerical series used by network protocol implementations and said that the compilation would be updated periodically. Its opening instruction directed developers who needed a link, socket, port, protocol, or network number to contact the numbers administrator for an assignment.

Coordination therefore began with a practical operation: select a value that would remain intelligible across independently built systems, associate it with a technical use, and publish enough common memory for other implementers to avoid collision.

That operation governed more than Internet network addresses. RFC 790 assembled network numbers, Internet version numbers, Internet protocol numbers, port or socket numbers, and link numbers. A protocol number told an implementation how to interpret the next level carried by Internet Protocol. A port or socket number identified an endpoint associated with a service. A link number named a protocol or purpose on the ARPANET Host/IMP interface. The common object was a technical position whose coordinated meaning depended on a maintained reference.

Network numbers made the structure especially visible. RFC 790 described a 32-bit Internet address divided into a network portion and a local portion. Class A used a 7-bit network number with a 24-bit local part; Class B used fourteen and sixteen bits; Class C used twenty-one and eight. These forms created differently sized administrative domains inside the same address architecture. The tables then attached names to selected network values and marked other portions as reserved or unassigned.

A class described address structure, an assignment associated a network identifier with a recognised network, and reservation withheld capacity from ordinary assignment. None of those acts required the register to collapse the three categories into a single notion of possession.

The version-number series shows that the register’s method was broader than address-allocation policy. Its field was four bits long, and values were attached to named versions, left unassigned, or reserved. The protocol-number field was eight bits and likewise joined recognised uses to numerical values. In both cases the coordinating problem was semantic stability. An implementation encountering a number needed a shared answer about what that number signified. The administrator made that answer portable across implementations by publishing an association and preserving unused positions for later action.

Technical scarcity existed wherever a field was finite, but RFC 790 expressed the current condition through assignment categories, not through an economic ordering of uses.

Port and socket numbers introduced an additional partition. RFC 790 described ranges for network-wide standard functions, host-specific functions, future use, and experimental functions, then listed particular services against particular values. The partition protected standardisation and experimentation within one namespace. A reserved range was an active administrative choice because it preserved capacity for a stated purpose; an unassigned value remained open to a future assignment; a named port gave disparate hosts a common convention. The register thus recorded the disposition of numerical positions at publication time.

It offered a way to coordinate change without treating every position as equivalent in technical function or requiring a financial comparison among applications.

Link numbers reinforced this emphasis on recognised use. Their table included values for protocols, measurements, maintenance, experiments, and reserved purposes connected with a field in the ARPANET interface. By placing link numbers beside networks, versions, protocols, and sockets, RFC 790 defined assigned-number administration as a family of namespace tasks. The administrative action was consistent even when the underlying technical consequence varied: identify a numerical series, preserve uniqueness within it, attach selected values to functions, and leave the remaining state visible.

A later observer may care far more about IPv4 addresses than experimental link values, but the 1981 compilation itself was organised around coordinated numerical meaning.

The register’s sparseness was functional. A row was sufficient for implementers to recognise a value without narrating the applicant’s plans, the administrator’s internal process, or the downstream importance of the assignment. The tables instead made currently assigned values from several numerical series available as a periodically updated reference, placing public technical state at the centre of the document.

What the document affirmatively supplies is the published result: a set of numerical series, named technical associations, reservations, and open positions, maintained by an identified assignment function and intended for periodic revision.

Nine years later, RFC 1174 altered the scale at which that object was administered. In August 1990 the IAB recommended distributing the assignment of network and autonomous-system identifiers as Internet growth became faster and more international. The document kept IANA and the Internet Registry institutionally distinct. It described the IANA function at the USC Information Sciences Institute as holding authority for numerical identifiers required for Internet operation, and the Internet Registry at SRI International’s DDN-NIC as performing registration for network and autonomous-system numbers.

The proposal addressed workload and global reach without turning the registered object into a different kind of thing.

Block delegation was the key scaling device. RFC 1174 proposed that the Internet Registry allocate blocks of network and autonomous-system numbers to organisations approved by the Coordinating Committee for Intercontinental Research Networking. Under the proposal, approved organisations would exercise further assignment authority within those blocks. This nested arrangement would have changed the granularity of administration: a central function allocated a defined pool to a delegated registry, and that registry handled assignments drawn from it. The design distributed work and placed decisions closer to regional settings.

The globally coordinated object remained assignment state inside a bounded numerical space, because each delegated act still had to fit an aggregate account of unique identifiers.

The proposal also preserved a default route. In an area without an identified delegated authority, the Internet Registry was recommended to serve as the registry, at least initially. That clause joined regional expansion to continuity of access to the assignment function. Delegation was conditional on institutional capacity and approval, not a declaration that every region already possessed an operating registry. Candidate registries were expected to meet IANA and the Internet Registry to review procedures and requirements, and the resulting distributed mode was to be documented in later RFCs.

The recommended structure connected authority, blocks, institutional preparation, and shared procedure.

The recommended information architecture was designed to hold the distributed arrangement together. RFC 1174 assigned the Internet Registry continuing responsibility for the database of registered networks and autonomous systems, along with the list of root Domain Name System servers. It recommended copies of aggregate registration databases for delegated registries to improve redundancy and access. Updates, however, were to remain centralised at the Internet Registry, after which complete copies were redistributed in a timely fashion by file transfer or other means. Regional access and central updating were separate functions.

Replication broadened availability of the common record; a single update point protected coherence in the combined administrative state.

The IAB’s scarcity language sharpened the reason for care. RFC 1174 called Class A and Class B network identifiers an “increasingly scarce commodity” and said allocation required thoughtful attention. That statement concerned classes whose supply was constrained by the address architecture and prior assignment. It supplied no transaction series or unit of monetary comparison. Its immediate institutional expression was delegated blocks, retained central functions, aggregate databases, and procedural coordination.

Scarcity entered policy as a condition bearing on allocation judgment, even though the object carried through the recommended machinery was still the state of identifiers and registered networks.

Later scholarship helps separate categories that the period records address unevenly. The Richter, Allman, Bush, and Paxson primer, prepared in 2014 and associated with a 2015 publication, distinguishes allocation from actual use, routing, and market exchange. That separation prevents an assigned entry from standing in for four claims at once. RFC 790 establishes a published assignment state; alone it shows nothing about whether every part of an assigned block was used or routed. RFC 1174 establishes a recommended allocation architecture and a scarcity warning; alone it establishes no later market transaction.

The primer is retrospective scaffolding for analysis, not evidence of what 1981 or 1990 actors knew.

Seen through that separation, the early administrative object was precise but limited. A registry was able to preserve uniqueness, record a network, distribute a block, replicate an aggregate database, and centralise updates without creating a record of economic exchange. Assignment identified an authorised technical association within the coordination system. Use concerned deployment inside the assigned space. Routing concerned reachability through operational networks. A market concerned exchange under terms and prices.

Keeping these objects apart makes the historical achievement clearer: the early system built an authoritative memory for numerical coordination before market categories became part of its routine public record.

Making Registration a Federal Service

When NSF solicited Network Information Services Managers in the spring of 1992, it translated parts of network coordination into procurable work. Project Solicitation NSF 92-24 divided the programme into Registration Services, Directory and Database Services, and Information Services. NSF selected a separate organisation for each award area: Network Solutions for registration, AT&T for directory and database work, and General Atomics for information services. The three providers retained distinct contractual responsibilities, yet NSF required collaboration so that users encountered a seamless InterNIC interface.

The procurement made institutional boundaries explicit around functions that users were supposed to experience as one service environment.

The award structure connected a technical registry object to people, systems, schedules, and money. Network Solutions’ assignment covered non-military Internet users and networks receiving services previously provided by the Defense Information Systems Agency Network Information Center. The agreement tied performance to the federal solicitation, Network Solutions’ September 1992 proposal, its supplemental proposal of 19 October 1992, and the technical terms incorporated into the award. Registration was no longer described solely through a published result in a number table.

NSF specified an operating organisation responsible for delivering a defined service under federal support and supervision.

The NSF–Network Solutions Cooperative Agreement NCR-9218742 became effective on 1 January 1993. Its calendar separated four stages that must remain distinct: legal effectiveness on that date; a three-month phase-in; five years of operational support beginning 1 April 1993; and six months of no-additional-cost flexibility through 30 September 1998. The sequence gave transition its own period before supported operations began. The final flexibility date described the agreement’s scheduling envelope. It adds no affirmative observation about registration practice after the calendar-1995 endpoint used here.

The cooperative agreement was cost-plus-fixed-fee, and its cover gave an estimated total award amount of US$4,219,339. The estimate belonged to the supported registration-service arrangement. Compensation covered allowable direct and indirect costs plus a fixed fee under the award terms, with funding supplied through the mechanisms stated in the agreement. The figure was not final expenditure. Its institutional subject was Network Solutions’ performance of contracted work, including transition, operations, management, and reporting. NSF was financing a provider to administer services; the numerical pool was not an inventory purchased from that provider.

Article 3 identified the initial work. Network Solutions was to perform domain-name registration, domain-name-server registration, network-number assignment, and autonomous-system-number assignment. It was to provide registration according to RFC 1174 and, in cooperation with IANA, the Internet Registry was permitted to establish delegated registries for specified domains. That clause carried the 1990 distribution architecture into a service specification without merging IANA with Network Solutions.

IANA remained a distinct function in the coordination setting; Network Solutions bore duties under its own federal agreement; NSF defined and supervised the award.

The task list shows how the earlier administrative object became a service output. A network-number assignment changed registered technical state. Domain-name and name-server registrations added records needed by the naming environment. Autonomous-system-number assignment extended the unique-number function to interdomain administration. Each output required intake, processing, updating, and delivery of information. The procurement attached labour and performance terms to those acts. It left the successful assignment outside the category of commodities bought by NSF.

The federal purchaser paid for the capacity and conduct of registration across specified categories.

Article 4 converted timeliness into an observable service obligation. The agreement set three working days for a Class C request, five working days for Class B, and twenty-two working days for Class A. The interval began upon receipt of a completed template together with specifically requested information about network topology and prior address-space use, and ended with assignment of a number. These class-specific periods tied an administrative event to a clock. They created a performance question about processing completed requests under stated conditions, not an economic ordering of the blocks being assigned.

Availability was another dimension of performance. Network Solutions had to provide registration data to AT&T’s InterNIC Directory and Database Services operation. The clause connected the registration provider’s output to the separately awarded information function. NSF’s seamless-interface requirement therefore rested on collaboration across distinct providers: one produced and maintained registration information, another made directory and database services available, and General Atomics supplied the information-services area.

Institutional coordination occurred through assigned responsibilities and shared delivery expectations, not by treating “InterNIC” as a single actor with undifferentiated authority.

Quality measures accompanied turnaround. The agreement referred to formulae in Section J of Network Solutions’ revised September 1992 proposal and required them to be considered with the specified periods. This made quality, timeliness, and availability named dimensions of the service. Article 6 gave Network Solutions primary responsibility for quality, timely performance, and effective management. NSF retained support planning, oversight, monitoring, and evaluation.

The design allocated operational responsibility to the awardee and supervisory responsibility to the funder, establishing who had to produce performance and who reviewed the supported programme.

Flexibility entered through Article 5. The listed services were estimates of immediate and longer-term requirements for planning, and NSF reserved authority to increase, decrease, or modify their quantity, quality, content, or nature as needs changed. A modification called for negotiation and incorporation of corresponding adjustments to estimated costs, fees, and funding schedules. The variable exposed to renegotiation was the scope of service work. This clause allowed the procurement to respond to uncertain demand without presenting network numbers as units priced under each change.

The same article scheduled a project review by 31 December 1994 to consider continued funding and general direction for the remaining support period. The agreement thereby placed a formal decision point inside the award’s time structure. Contractual review concerned progress, support, and future programme direction. The clause establishes the design, not the outcome or actual performance: continuation was subject to an identified federal consideration, and the service provider operated within a framework that anticipated revision and supervision.

Article 8 made support incremental. Funding for later periods depended on review of progress, the proposed programme plan, and the availability of funds; additional allotments were to be made in stages. Reallocations above the agreement’s threshold required prior approval. This structure joined programme progress to future financing without turning the full estimated amount into an automatic statement of money spent. The award created an operating horizon, yet funds remained connected to periodic plans and federal decisions.

Registration capacity was procured through controlled support instead of a one-time purchase of a finished registry product.

Article 9 imposed an annual cycle. By 31 December, Network Solutions had to submit an Annual Report, Programme Plan, and Budget electronically and in ten hard copies. The submission needed goals and objectives expressed through technical criteria, milestones, and timetables sufficient to measure progress. Once approved, the plan became the basis for performance goals and funding for the next twelve-month operational period beginning on 1 April. Planning and retrospective reporting were thus linked: the provider stated intended work, NSF approved a basis for support, and later reports described performance against the programme.

The annual materials had to identify goals achieved, exceeded, or missed and explain significant deviations from the previous plan. They also covered educational accomplishments, innovations, external funding, income, and contributions. Actual line charges and expenditure had to be shown for the year and cumulatively, by functional area and for the project overall. These requirements made the cost of administration visible inside a provider-specific account.

They also enabled NSF to relate stated objectives to spending categories and programme changes, giving registration service a financial and managerial form absent from a bare assignment table.

Article 10 added a faster reporting rhythm. Significant problems affecting schedule or progress required timely notification. Weekly status reviews applied during phase-in until NSF directed otherwise. Quarterly reports covered major events, work performed, technical status, accomplishments, problems, collaboration, changed plans, pending approvals, expenditures, and line charges. A final report had to describe the work and difficulties with enough detail to permit replication by a reasonably knowledgeable party or organisation.

Reporting was part of the deliverable: NSF procured not only operational acts but also a continuing account of how the supported service was being conducted.

The agreement anticipated alternative registration or numbering systems and a user-based fee structure, subject to express direction or approval from the NSF programme official. Article 15 specified the treatment of registration-fee income if charging was authorised. Collected fees entered an interest-bearing account and were applied in order to costs created by imposing the fees, Network Solutions’ expenses charged to NSF, and collaborator expenses charged under their awards. Programme plans and later funding requests had to reflect that income. User payments, if introduced, were governed as project finance within the service arrangement.

Here the service-cost/resource-value difference appears in its institutional form. Administrative-cost finance paid for staff, systems, assignment work, information delivery, planning, and reporting; a scarcity price would use payment to allocate a limited numerical resource. The agreement specified the first and reserved federal control over fee income, without establishing the second. Registration was a procured service, and address assignment remained its technical-administrative output.

Regional Cost Recovery in Institutional Accounts

RIPE-139, the RIPE NCC Annual Report 1995, was published in 1996 and described the completed calendar year 1995. Its institutional subject was RIPE NCC, a regional service organisation separate from the IAB, IANA and Internet Registry functions, NSF, Network Solutions, AT&T, General Atomics, and Geoff Huston. The report made regional registration finance visible after the year had closed. That perspective differs fundamentally from a federal agreement setting prospective terms for a contractor: RIPE NCC classified receipts, expenditure, functional cost, reserves, and a stated next-year fee consequence within its own annual account.

The report identified 818 kECU as income received for services provided during calendar year 1995. The wording ties the inflow to service and period. It was money received by RIPE NCC for work provided in that completed year, recorded in the organisation’s financial account. The category demonstrates that a regional registry service had a defined income stream and that the annual report placed the stream before its readers. It says nothing about a sale of IPv4 holdings. Its useful comparison is with other calendar-1995 institutional categories recorded by RIPE NCC, because those categories share an organisation and reporting period.

Total expenditure for the same organisation and year was 535 kECU. This category captured RIPE NCC’s organisation-wide use of funds during calendar 1995. It was broader than registration alone and therefore defined the financial scale against which the annual report related income. In contrast with an estimated future award, the figure appeared in a report for a completed year. Even so, completed annual expenditure is not a finding about the performance of an individual allocation. It records institutional resources consumed across activities grouped by the organisation in its accounts.

Within total expenditure, the report classified 326 kECU as registration-services expense. Naming the function separately made registration’s operating cost visible inside the broader institution. The line was a component of the total, not a second statement of the whole cost base. It shows a regional organisation distinguishing the expense of a core service from its other expenditure. That classification is institutionally significant because the public number record needed no staff-cost or overhead fields for those costs to become legible elsewhere. The annual account supplied a financial view of the service function at a defined level.

RIPE NCC stated that income received for 1995 services equalled 153 percent of expenditure. The numerator was the organisation’s calendar-1995 service income, and the denominator was its reported expenditure for that calendar year. Transferring the numerator or denominator to a count of addresses, assigned blocks, applications, or routed space creates an unsupported ratio, because none of those technical populations defined the reported relationship. The controlled accounts support comparison among RIPE NCC’s period-aligned institutional categories. They do not yield a unit address price, recipient benefit, allocation efficiency, or value for the remaining pool.

The annual position permitted reserves, according to the report. Reserve creation was a treatment of institutional funds arising from the relationship the organisation reported between annual income and expenditure. It belonged to RIPE NCC’s capacity to carry financial resources within its service organisation. The supported act is the organisation’s treatment of the annual position as a basis for reserves. The report keeps those reserves separate from administered address value.

RIPE-139 also connected the reported position to lower fees for 1996. Those were charges for RIPE NCC services in the next calendar period. The decision shows an organisation using annual financial information to set the funding terms of its operations: money received, total spending, a functional registration cost, and reserves informed a stated service-fee change. Lower service fees are fully compatible with an increasingly constrained address pool because the two categories respond to different institutional variables. One concerns financing the provider; the other concerns allocation from finite technical capacity.

Staffing was part of the service organisation documented in the annual report, adding an operational dimension to its income and expense categories. Registration required personnel, internal organisation, and funded activity; it was not costless merely because the assigned resource lacked a scarcity charge. RIPE NCC’s published account made those inputs available through the categories chosen for its own finances. The account’s granularity also shows that “free address space” never served as a complete description of the surrounding institution. Personnel and facilities still had to be financed through an organisational mechanism.

The federal agreement and the regional report reveal two distinct forms of cost recovery and accountability. NSF specified ex ante duties for Network Solutions: a cost-plus-fixed-fee award, annual plans, budgets, milestones, incremental allotments, controlled fee income, and frequent reports. RIPE NCC described ex post finances for its own completed service year: income, total expenditure, registration expense, reserves, and a later fee decision. One instrument governed a funder-provider relationship; the other presented a regional organisation’s internal financial categories to its constituency.

Their shared subject was funded administration, not a shared chain of authority.

Time changes the kind of question each instrument supports. The cooperative agreement shows what Network Solutions was required to plan, perform, report, and finance from the effective date onward; it supplies requirements instead of results. RIPE-139 shows what RIPE NCC reported for calendar 1995 and supplies no retrospective account of federal-award performance. The comparison is strongest when institutional forms stay separate: prospective procurement sets obligations and decision points; completed accounts classify a year’s receipts and spending. Both make service cost visible through their own denominators.

Their treatment of fees also differs. Under the federal award, a user-based fee structure required the NSF programme official’s direction or approval, and any resulting registration-fee income followed an ordered project-expense rule. RIPE NCC reported a service-finance position that permitted lower fees in the following year. In the first case, fee income was a contingent and controlled input to a cooperative project. In the second, fees were part of an operating organisation’s reported annual funding cycle. Neither instrument identified a payment as a market-clearing amount for address scarcity.

These forms of accountability were concrete without being interchangeable. NSF’s terms made Network Solutions answerable for programme plans, technical criteria, schedule reports, spending lines, and management under a named award. RIPE NCC’s annual report made categories of its service economy visible for a completed period. The agreement assigned oversight roles; the annual report presented organisational finance. No common conclusion about institutional success follows simply from the presence of either mechanism. Their evidence concerns the structures through which money and administrative work were made legible.

The regional account therefore changes the history in a specific way. By the end of the affirmative 1981–1995 period, published records were able to identify a regional registry’s service income, organisation-wide expenditure, registration-specific expense, reserve treatment, and next-period service fee. Those categories were considerably richer than a number table as financial descriptions of an institution. They remained categories about RIPE NCC itself. The administered resource appeared through registration activity, not as an asset whose value was added to the organisation’s accounts.

This asymmetry was not an accounting error demonstrated by the report. It followed from the different objects selected by the two record systems. RIPE NCC’s annual account classified the organisation’s operations; registry records classified identifiers and assignments. A complete claim must respect those objects. The accounts sustain a history of financed regional coordination and disclosed cost categories. They do not sustain a history of address sales, recognised resource assets, distributable profit, or verified downstream effects of reserve policy.

Institutional finance was visible because RIPE NCC chose categories fitted to service administration.

Pricing the Scarcity Problem

By August 1990, the IAB had already placed scarcity inside identifier policy. RFC 1174 described Class A and Class B network identifiers as an increasingly scarce commodity requiring thoughtful allocation. The warning arose in a recommendation about distributing assignment responsibilities under rapid growth and international expansion. It linked finite classes to judgment and institutional scale, yet expressed no pricing rule. The recommended response lay in delegated blocks, approved regional registries, a default Internet Registry, retained IANA and central registry functions, and an aggregate information system.

Scarcity first entered this account as a reason to govern allocation more deliberately.

December 1994 brought a different kind of intervention. In RFC 1744, Geoff Huston analysed management of the globally unique 32-bit IPv4 address space as the unallocated pool shrank. He argued that the deployed Internet gave address space “considerable market value” and examined parameters including efficiency, availability of technology, availability of service providers, and market size. The claim was a contemporaneous economic proposition about finite capacity and demand. It came from Huston, not from RIPE NCC’s service accounts, NSF’s procurement terms, or an undifferentiated registry community.

Huston decomposed the prevailing allocation model into first-come-first-served, once-and-for-all, and free: FCFS, OAFA, and FREE. Each term identified a different decision variable. FCFS governed priority through arrival order. OAFA described the duration or continuing character of an allocation. FREE described the resource-price term. The decomposition matters because an allocation regime is not defined by price alone. A queue can distribute opportunity across time, a duration rule can shape later availability, and a zero resource charge can affect demand. The three features may interact, but each requires its own institutional analysis.

The three variables also assigned different responsibilities. A registry operating FCFS needed a rule for ordering completed requests. An OAFA arrangement made the original allocation consequential for later availability because no routine term reset the decision. A resource price, if adopted, required a basis for quantity and duration. Huston’s framework thus moved debate from the generic word “free” to the institutional levers that structure access.

FCFS offered an administratively intelligible ordering rule. A completed request entered a sequence, and arrival order provided a visible basis for priority without comparing every applicant’s prospective economic return. Under a finite pool, Huston argued, sequence also placed earlier applicants in a different position from later ones because prior allocations reduced what remained. That is an incentive proposition about chronological access. It should not be converted into a finding about any named application or an observed distributional outcome, since RFC 1744 contains no case series from which such a result can be calculated.

OAFA directed attention to time after assignment. Under Huston’s characterisation, an allocation was made once instead of moving through a regular reallocation cycle. Duration changes the future pool even when the initial priority rule stays constant. A continuing assignment, a renewable term, and a scheduled return mechanism place different demands on recipients and administrators. Huston used the once-and-for-all label to argue that the allocation decision had consequences beyond the day of issuance. His text raised this temporal structure as part of the economic problem and left the legal form of recipients’ interests unsettled.

FREE isolated the amount charged for the resource. Registry administration still consumed staff time, systems, and funding, as the federal service agreement and RIPE NCC accounts demonstrate. Huston’s category addressed the lack of a resource price reflecting scarcity, not the absence of every administrative expense. This kept two questions separate: how an institution finances processing and record maintenance, and whether an amount influences access to finite address space. The first can be answered through budgets or service fees. The second requires a pricing rule tied to quantity, duration, return, or demand for the limited pool.

Huston expected future exchange to make competing valuations more visible. He anticipated that entities with inefficiently used blocks might transact with parties seeking address space after exhaustion of the unallocated pool, and he discussed trading as a response to demand that primary free allocation was unable to satisfy. He also forecast hoarding: the combined prospect of future scarcity, continuing assignments, and free initial allocation might reward obtaining or retaining more space before the pool closed. These mechanisms were propositions about incentives and future conduct, attached to the architecture he described.

Concentration formed another part of the forecast. Huston considered the possibility that large service providers with substantial holdings and established positions might dominate trading, influence prices, create entry barriers, or control recording functions needed to validate transactions. Leasing appeared as a possible arrangement under which address use might be made time-bound and renewal policies might affect conservation. Charges scaled to larger blocks, together with an opportunity cost for unused space, were presented as possible tools for encouraging more economical use and extending the unallocated pool’s life.

Forecast and observation enter only once in evaluating these propositions: RFC 1744 records Huston’s December 1994 predictions and proposals, not completed transaction volumes, observed prices, verified hoarding, demonstrated monopoly, adopted leasing, or measured conservation results. It also stated that no explicit IPv4 pricing policy had successfully entered the allocation process by that date. The historical fact is that these mechanisms had become articulable inside a contemporaneous RFC. Their later accuracy, adoption, or effect requires different evidence and is not imported into the 1981–1995 period.

Pricing itself may pursue different objectives. Huston distinguished an amount designed to recover administrative cost from one designed to express scarcity. A cost-recovery charge begins with the resources needed to receive applications, assess requests, update records, and operate the registry. A scarcity charge begins with limited supply and uses price to shape how much is sought, how long it remains assigned, or whether it returns. The same payment channel might collect either kind of amount, so institutional purpose and formula determine its meaning. A fee label alone reveals nothing about which allocation problem it addresses.

The proposal also exposed an authority problem. Huston discussed international recognition, control of the address pool, and the difficulty of reaching consensus around a common price mechanism. Pricing globally unique identifiers was not simply a matter of adding a charge to existing forms. It required an institution able to set the rule, recognise transactions or time limits, preserve an authoritative record, and command sufficient acceptance across the network environment. Any such mechanism also needed a record able to distinguish primary allocation from later exchange and connect the transaction to globally unique assignment state.

That record requirement connected exchange design back to the earlier coordination object. Any price mechanism needed an authoritative account of which block remained available, which assignment continued, and which recorded change returned or transferred capacity. Economic incentives depended on technical uniqueness; without coordinated state, a scarcity payment alone offered no reliable account of the resource affected.

A compact legal boundary keeps the economic analysis in place. Huston’s discussion of ownership, lease forms, and transaction recording was part of his proposal; the period evidence here leaves title, collateral, transfer, portability, continuity, liability, remedies, and review unsettled for actual recipients. Allocation, use, routing, and exchange remain analytically separate, as the later primer makes especially clear. An allocation entry states an administrative relation. Alone it discloses nothing about how the space was deployed, announced, valued, or treated under law.

The access counterargument deserves equal institutional weight. Free issuance lowers the immediate monetary barrier to requesting number space, especially for organisations with limited capital or uncertain demand. A thin technical process can also move faster and impose fewer information costs than auctions, bargaining, recurring valuation, or transfer review. Those advantages are not Huston’s attributed position in this evidence.

They are countervailing mechanisms supported by the design of technical coordination: neutral uniqueness and broad participation may be served by refusing to make willingness to pay the primary allocation criterion.

RFC 1174 and RFC 1744 therefore frame different responses to a finite administrative space. The IAB recommendation used institutional distribution, central coordination, and thoughtful allocation to handle growth and scarcity. Huston analysed priority, duration, and price, then proposed incentives that might alter demand and retention. The later primer prevents these moves from being mistaken for observed use, routing, or exchange. By the end of 1994, the value proposition was no longer confined to whether registries cost money.

It concerned which variable—need, queue position, time, payment, or return—should govern access to a diminishing pool.

Accountability Around an Unpriced Administrative Object

Accountability begins with a choice of object. RFC 790 made assigned and reserved values visible. RFC 1174 added delegated blocks, registered-network information, and central updates. The NSF agreement exposed service tasks, providers, schedules, plans, and spending lines. RIPE-139 exposed regional income and expenditure. RFC 1744 made queue priority, assignment duration, and scarcity price separate economic questions. Each instrument rendered a domain governable by naming what administrators or readers were able to inspect.

The first accountability consequence concerns the quality of the common technical state. If uniqueness is the object, a registry must maintain coherent assignments across numerical series and organisational boundaries. RFC 1174’s combination of block delegation, aggregate database copies, and central updating addressed this requirement at scale. Delegated registries needed access to the combined information, yet changes entered through a central update function before redistribution.

The instrument design therefore supports questions about coordination: which function assigns blocks, where updates enter, how copies reach regional actors, and whether a default registry remains available where delegation has not been established.

A second consequence concerns the boundary of authority. RFC 1174 distinguished IANA’s broader identifier function from the Internet Registry’s role for network and autonomous-system numbers. The InterNIC award then distinguished NSF as funder and overseer, Network Solutions as Registration Services provider, AT&T and General Atomics as providers of other service areas, and IANA as a cooperating function in delegated registration. Accountability improves when these roles stay named.

A seamless user interface left those duties separate, and an aggregate registry function left policy, finance, and economic analysis with their actual authors.

Procurement created a third set of questions. Because NSF defined registration as a service, it specified what counted as delivery. Domain, name-server, network-number, and autonomous-system-number tasks formed the initial scope. Class-specific turnaround periods gave timeliness an operational form. Availability required registration data to reach the directory and database provider. Quality formulae, programme plans, annual budgets, quarterly reports, and problem notices created additional observation points.

These clauses support inquiries into required conduct and reporting, without supplying results that the agreement itself merely anticipated.

Funding design added control over change. NSF reserved authority to alter the quantity, quality, content, or nature of services and required corresponding adjustments to costs, fees, and schedules. Continued support depended on progress, future plans, available funds, and incremental allotments. A scheduled project review created a point for general direction on later support. Fee income, if authorised, followed an ordered expense rule and had to appear in later plans. The accountable object was the federally supported programme: its scope, performance basis, money, and adaptation over time.

Regional accounts created a different accountability surface. RIPE NCC classified service receipts, organisation-wide spending, registration-specific expense, reserves, and later service charges for a completed year. Readers were able to relate a functional cost to the institution’s total expenditure and see how the report described its annual financial position. That visibility supports questions about the financing of regional operations and the categories used to present them. It establishes no reader approval or institutional legitimacy. Publication makes the account inspectable; it cannot predetermine the judgment readers should form.

Scarcity changes the questions asked of assignment design. RFC 1174’s policy called for thoughtful allocation of increasingly constrained Class A and Class B identifiers. That language directs attention to how a delegated structure preserves consistency and how judgment operates inside block assignment. RFC 1744 adds priority and duration: who enters first, what continuing effect follows, and which incentives arise when the resource-price term is zero. These are consequences of instrument design because a queue, an assignment term, and a fee rule structure future choices even before an exchange market supplies a monetary signal.

Neutral uniqueness is a serious answer to those concerns. A technical register offers a common status independent of each participant’s wealth, accounting practice, or bargaining power. Arrival order is observable and comparatively inexpensive to administer. Central updating reduces the risk of incompatible local records. Delegation can expand processing capacity and regional access without forcing every request into an economic contest. These mechanisms place interoperability first, and the period instruments show why that priority had institutional substance rather than serving as a rhetorical cover.

Speed is another substantive counterargument. The InterNIC agreement’s three class-specific turnaround periods treated delay as a performance concern. Adding auctions, negotiated prices, repeated appraisals, or elaborate transfer assessment would require more information and administrative time. The sources provide no comparison of actual outcomes under those alternatives, but the contract establishes that timely processing itself was an intended service quality. A governance design that values rapid network expansion has reason to count the burden imposed by any mechanism layered onto assignment.

Low transaction costs extend that point. RFC 790’s published states allowed implementers to consult a common record without reconstructing the economics of every entry. RFC 1174 scaled administration through blocks and replicated data, tools suited to a growing international network. A procedure limited to technical need and unique assignment asks less of applicants and administrators than a system that also establishes bids, valuations, time-limited interests, and exchange conditions.

Reduced information demand can be valuable where institutional capacity and shared economic rules are limited, even though the evidence assigns no such motive to a named decision-maker.

Lower monetary barriers complete the countercase. If number resources carry no scarcity charge, an applicant need not outbid better-capitalised organisations merely to enter the allocation process. Service finance may come from a funder or from charges designed around operating cost. This arrangement separates the ability to pay for administrative infrastructure from the criterion used to distribute identifiers. The benefit is potential access on non-market terms; the cost identified by Huston is that demand and retention may remain weakly disciplined by price. The instruments define the tradeoff without ranking its effects.

The absence of a routine resource valuation also places limits on financial accountability. A number register gives no monetary amount to reconcile against institutional income, and a service account gives no resource price to compare with the assignment state. The appropriate response is not to manufacture one by joining unlike records. It is to ask separate questions: whether the technical state is coherent, whether the service performs under its stated terms, whether organisational funds follow their rules, and whether the allocation policy addresses scarcity through criteria other than price.

Each question has an evidentiary object fitted to it.

Economic accountability can still arise without a market. RFC 1174’s “thoughtful allocation” invites scrutiny of the criteria applied to constrained classes. RFC 1744’s taxonomy invites scrutiny of the queue, the duration of assignments, and the incentives created by a zero resource charge. Block delegation invites scrutiny of how authority is distributed and how the aggregate record remains consistent. A service contract invites scrutiny of timeliness and programme finance. None requires treating an identifier as private inventory; all require clarity about the decision rule actually being administered.

Institutional consequences therefore arise from selection, not from a single collective decision. The IAB selected a recommended architecture for distributed assignment under retained central functions. NSF selected a procurement form and divided the InterNIC service areas among three providers. Network Solutions accepted duties specified in its cooperative agreement. RIPE NCC selected categories through which to report its calendar-1995 finances. Huston selected economic variables for an argument about scarcity. Their actions do not form one institutional mind or a demonstrated chain of learning.

Comparison reveals coexistence among different governing objects.

That coexistence leaves a durable analytical question. A record can be perfectly suited to maintaining unique technical state and still provide little information about scarcity incentives. A contract can expose costs and performance obligations without valuing the resource processed through the service. An annual account can explain how an organisation is funded without pricing each assignment. An economic proposal can identify possible incentives without supplying observed outcomes.

Accountability depends on recognizing which decision each instrument can discipline, then resisting claims that require an object the instrument never undertook to record.

The practical consequence is a layered governance agenda grounded in the period documents. Preserve coherent assignment state; keep authority and update paths identifiable; specify service tasks, timing, and funding; report institutional income and expense in period-aligned categories; and make allocation criteria intelligible when scarcity bears on choice. Price is one possible mechanism in Huston’s analysis, not the sole form of accountability. Speed, neutral uniqueness, reduced transaction cost, and lower monetary barriers remain valid considerations alongside conservation and future access.

Conclusion: The Administrative Object and Its Limits

A registry’s governing power begins with what it chooses to make legible. Early assigned-number records made coordinated technical state legible: a value belonged to a named series, carried a recognised use or reservation, and occupied a position that others could consult. Distributed administration extended that object through delegated blocks, shared aggregate information, and a controlled update path. These choices created a neutral reference across organisations, an achievement whose value did not depend on attaching a monetary amount to each entry.

Funding instruments selected a different object. They made the production of registration service legible through duties, time, staffing, plans, expenditure, income, and supervisory roles. That accounting supported operational control because its categories matched the institution being financed. By itself, it answered none of the questions about distributing finite number space. The same restraint applies in reverse: assignment state supports uniqueness and availability decisions, yet supplies no institutional service account.

Scarcity policy and economic argument expanded the field of decision without erasing those boundaries. A finite pool raises questions about priority, duration, conservation, access, and future demand. Price may address some of them; thoughtful allocation, delegated judgment, or technical change may address others. Speed, low transaction costs, neutral uniqueness, and lower monetary barriers remain substantive considerations because the coordination system had to admit participants and preserve interoperability as it allocated constrained capacity.

The new synthesis is therefore about institutional fit. A record sustains the claims encoded by its administrative object: technical registers sustain claims about published assignment state; contracts sustain claims about required service; completed accounts sustain claims about organisational finance; policy and analytical texts sustain claims about recommended structures and argued incentives. Governance becomes more exact when those capacities are joined at the level of decision, not fused into a fictional common account.

The early record supports that disciplined plurality—and no broader conclusion about value than its chosen objects can bear.