RFC 1744

Okay bestie, the title block first 🎀. This is RFC 1744, Observations on the Management of the Internet Address Space. It's from the Network Working Group, written by G. Huston at AARNet, and it's marked Category: Informational (so it's giving info, not a standard 💅). It's dated December 1994, and it doesn't obsolete or update anything.

Status of this Memo 📜

Okay, this part is the legal-ish stuff, but it matters 🫶. This memo is here to give information to the Internet community 📚. It does not specify an Internet standard of any kind, so no cap, it's not a rulebook. You can distribute it with no limits 🌸.

Abstract 🔮

This memo takes a look at the issues with how the IPv4 address space is managed right now 🧐, and what could happen to all of it as the pool of unallocated addresses keeps shrinking 📉. It checks out possible changes to those management practices and what outcomes they might lead to 🤔. Then it pulls out some general conclusions ✨ and notes why they matter for making address management policies for IPv6 🦋.

1. Introduction 💖

Let's get into it 💅. The thing we're actually looking at is the allocatable, globally unique IPv4 address space 🌍. That covers the address groups handed out from one big comprehensive pool to specific entities. Those entities are then free to assign individual address values inside their group to individual hosts 💻. The technology treats that whole group as one single network entity 🧩.

Right now, addresses get handed out on a freely available, first-come, first-served basis 🏃‍♀️. There are also a bunch of administrative rules that try to steer things so the space gets used sensibly 🧠. They also aim for some fairness in availability, sort of by splitting things into multi-national "regions" [1] 🌏.

If we dig into the current management policies, it's useful to know that IPv4 addresses share a bunch of traits with other public space resources 🌳. An economic look at them shows some parallels, including:

• the finite nature of the resource: It's finite because of the underlying tech, which defines addressed entities with a 32 bit address value 🔢. The whole pool is 2**32 distinct values, and not all of them can be given to end systems 🙈.
• the address space has considerable market value: It's valuable because the Internet tech is widely deployed, which lets uniquely addressed entities do direct end-to-end transactions with their peers over the Internet 🤝. That value is also shaped by how efficiently the allocated space gets used, whether end system based Internet technologies are available, whether Internet-based service providers are around, and how big the resulting Internet market is 📈.
• address space management is a necessary activity: Management processes are required to make sure every allocation is unique and access stays fair 🧹. Someone also has to keep the allocation record databases maintained, over and over 🔁.

Addresses have been getting handed out faster and faster in recent years 🚀, so IPv4 is now clearly a finite resource 😭. Current projections, assuming today's demand keeps going, say the unallocated pool runs out in the next 6 - 12 years (rephrasing current interim projections from the IETF Address Lifetime Expectancy Working Group) ⏳.

That raises two follow-up questions 🤔. First, what's the likely outcome if the unallocated pool does run out? Second, are there corrective processes that could be applied to the current address management mechanisms, so allocation is more equitable and the life of the unallocated pool could be stretched out? 🧐 Those two issues get covered in the following sections 🎀.

(Part 1 of 4, more coming bestie 💜🧚)

2. Outcomes of Unallocated Address Space Exhaustion - No change in current Address Management Policies ✨

okay so as the unallocated pool starts running low, the first thing that breaks is its ability to handle big block requests 😭. Those requests are already coming in from utility operators, and honestly, huge blocks are probably going to keep being a thing in address pool management 💅.

And here's the tea: most of the allocated space is being used super inefficiently right now. RFC 1466 notes efficiency figures under 1%, and way better utilisation is totally doable with newer routing tech like Variable Length Subnet Masks (VLSM) and disjoint subnet routing 🎀. Since the pool keeps shrinking and not everyone can get a block, the people holding allocated space and the people who want it will probably start making private deals 🤝. Say someone has a massive but underused block. They could sell it to a third party and then ask for a smaller block from whatever's left in the unallocated pool. Both blocks would end up used more efficiently, but the one who wins biggest is the entity holding the large blocks, basically cashing in on the opportunity cost of better efficiency 💰.

These resale and trading deals that dip into the unallocated pool would most likely be short-lived 🌸. The high returns would crank up pressure on the pool, people would race to grab large blocks for later resale once those blocks are gone, and depletion would speed up even more 🔥.

Once the unallocated pool is totally exhausted, a free trading market in address blocks is a probable outcome, with blocks bought and sold between trading entities 🛍️. If it's unregulated, the market would price address space around the common expectation of its value, reflecting demand, the opportunity cost of more efficient address use, and the opportunity cost of deploying additional or alternate internetworking technologies to IPv4. Now this part is iconic: the registry (or whatever takes its place) becomes basically a title office 📜. It records transactions to keep "ownership" accurate, so a buyer can check the seller's "title" to make sure the sale is legit. That means Internet address registries would start to look like "title offices", which are usually service entities that charge "lodgement fees" to fund recording title changes. Whether existing registries adapt or someone else steps in is a moot point. Either way, the function is a necessary part of this kind of trading environment 👑.

It's also expected that, without regulation, trading would quickly concentrate among major Internet providers, with a small number of entities controlling most of the volume (market efficiency means traders with big inventories would be more efficient in this space 🧠). ISPs would likely dominate this area because they have the most at stake in this resource. That would let an ISP plan service lifetimes with a lot more confidence, since it could set the price of the basic address resource and build an address pool as a hedge against local depletion for its client base 🫶. But there's a real risk too. These entities could naturally form a trading cartel and lock in a trading monopoly, putting up a formidable barrier that keeps new service providers out, which then leads to monopoly-based service pricing 🚨. It gets worse if the providers set up their own "title office". The major trading block would then control the only way to establish legitimacy of "ownership", and in terms of anti-competitive trading practices, that's a very seriously damaged outcome 💀.

If achieving much higher address utilisation efficiency is relatively cheap, then the resulting market is bounded only by the costs of renumbering agility 🦋. Renumbering would happen when a different address block is acquired because local address requirements change, and it might occur on cycles lasting anywhere from weeks to years. Markets would also be limited by deployment costs. Trading addresses inside a provider's own domain would barely affect deployment services, since the aggregated routing picture stays the same for the provider and its peers. Trading small blocks across provider domains, though, would drive up operational service costs because of higher routing costs, as efforts to create aggregated routing entries get frustrated by address leakage into other routing domains 🧩.

Looking at this whole environment, the major technical outcome is strong pressure for dynamic host address assignment services 🧚. Hosts connecting and disconnecting from the Internet would cause a local state change in allocated addresses, which might in turn trigger extended dynamic renumbering from time to time to match longer-term address usage trends. A strengthening market for dynamic address translation technologies is also reasonable to predict 🌈. That's an alternate client strategy to buying large blocks from the trading market: the client uses a private, potentially non-unique address space inside its network and dynamically translates end host addresses into a smaller unique Internet-routed address pool to support external end-to-end sessions. There's also a stronger market for firewall boundary technologies, which likewise allow private address space inside the client domain 🔥.

We can't predict specific outcomes accurately, but increasing overall efficiency of address utilisation will probably be most visible only after the unallocated address pool has been exhausted. At that point there's a strong economic motivation for it across the entire Internet address space 💜.

And as a cautionary note about evolutionary technologies for IPv4: they won't automatically get a quantum increase in economic viability just because the unallocated address pool is exhausted 🙈. They'll only gain an additional advantage over IPv4 once the marginal cost of deploying IPv4 address space more efficiently exceeds the marginal cost of deploying new technologies. That situation may not happen for some considerable time after unallocated address pool exhaustion ⭐.

3. Modification of Current Internet Address Management Policies 🎀

okay bestie, so right now the way addresses come out of the unallocated pool is "first come first served" (FCFS) 💅 and allocation "once and for all" (OAFA), which means once you get one it's yours forever 🫶. And there's no charge at all for getting one (FREE) 🍓. Main character energy for whoever shows up first ✨.

ngl here's the tea 🫖. When a finite resource gets handed out like this, you end up with a secondary market, where the stuff that was already allocated gets traded at its market valuation 💸. This secondary trade only really helps the entities who got a primary position from the pool 😭. And it's kind of concerning that the best move while the pool still exists is to hoard allocated addresses, because you're betting the secondary market will show up once the pool runs out 🙈. That market doesn't really help the original address management operation, and it doesn't necessarily help the wider community of current and potential people who care about the Internet either 💀.

Free allocation also means the big ISPs end up running address management 👑. Without end-client charging policies that can keep an independent address management function going, the entities with the most riding on the quality of address allocation and registration will fund it anyway, because nothing else is doing it. The risk is that putting the biggest asset of a communications medium into the hands of the companies already trading in that medium raises the chance of anti-competitive monopolistic trading practices 😬.

An alternate strategy is allocation and recovery 🔄. The allocation of an address is restricted to a defined period, so it's basically a lease 📜. In that kind of environment, pricing could be a tool for an efficient and dynamic address allocation mechanism 💖. Though it's immediately asserted that pricing alone may be insufficient for a fair, equitable and rational outcome of address accessibility and later use 🌸, so pricing and associated allocation policies would be a normative approach to this public resource management issue 🧚.

Pricing as part of a public resource management framework is a super common practice 🌈. Price and policy are used together to ensure equitable access, efficient utilisation, and availability for reallocation after use. Pricing that costs more for larger address blocks helps a diversity of entities who want allocations get access (basically a scarcity premium) 💅, and pricing can be devised to encourage provider-based dynamic address allocation and reallocation environments 🦋.

Just like a conventional lease, the lessee would get the first option to renew when the lease period ends, which lets the lease develop and hold a market value 🎀. Pricing policies like this would effectively imply a differential cost between deploying a uniquely addressed host with full Internet peering and reachability (including local reachability), and deploying a host with a locally defined (and potentially non-unique) address that's restricted to local reachability 🧚‍♀️.

It's also observed that pricing policies can encourage efficient address space utilisation through the opportunity cost of unused space 🙌. That's balanced by the potential cost of host renumbering practices, or the cost of deploying dynamic address allocation or translation technologies 🫧.

There are a number of anticipated outcomes of a management mechanism that includes pricing elements for the IPv4 address space 👀:

1. Firstly, current address space utilisation projections (the anticipated useful lifetime of the pool of unallocated addresses) would extend further into the future 📈. That's because of cost pressure for more efficient address utilisation, the additional cost of issuing a local resource with a globally unique address, and the opportunity cost of extravagant use of global addresses on purely local domains 💸.
2. Secondly, dynamic host address binding technologies and dynamic network address translation technologies would be expected to be widely deployed 🚀. That's because they'd look like cost opportunities compared with extensive static host address binding using globally unique addresses. Using these technologies would further extend the lifetime of the address pool ✨.

Pricing practices could be applied to all future address allocations, leaving entities with already allocated address blocks outside of the lease mechanism 🙈. Or, previous allocations could be converted to leases 👑. That way one management policy covers the entire address space, which gets the most out of pricing in terms of maximising the lifetime and the value of the address space 💎. In a situation like that, some level of recognition of previous implicit OAFA allocation policies can be offset by delaying conversion to a lease, and also by converting previously allocated addresses to leases while waiving the lease purchase costs in those cases 🫶.

4. Internet Environment Considerations ✨

Okay so, pricing IPv4 addresses as part of the overall address management framework is honestly not a new idea 🙈. Pricing policies could give us some real perks: better efficiency of use, fair and equitable access, security of allocation, a resulting market value, and some time offsets on when the address pool runs out. But it is undeniably the case that no explicit pricing policies have been successfully introduced into Internet address allocation processes to date 💀.

There are two big reasons this analysis gives for that. 🔥 The first is how uncertain it is where primary ownership of the IPv4 address space comes from, especially the unallocated address pool. The pool has been run under policies drafted by the Internet Assigned Numbers Authority (IANA). Those policies came out of the same consensus-seeking approach used in the Internet Standards process. In that kind of environment, unilateral declarations of ownership and related claims of policy control have a hard time playing an effective role in the Internet community, and they are generally unable to gather consensus support. (It can be argued that "ownership" isn't even a relevant concept here, because what really matters about these address elements is their uniqueness across the global domain. That uniqueness only works through shared recognition of a coordinated and reliable management environment, not through where the resource historically came from. 🧚) The second reason is that the address space isn't formally recognised as a shared international resource that falls under national public resource management policies and the administrative entities of each nation. It also isn't recognised as a private resource owned and run by a single entity. 🤷‍♀️

Recent policy changes have assigned large chunks of the unallocated pool to international bodies on a regional basis, and those bodies then assign pieces to bodies inside national contexts. 🌸 Throughout all of this, the allocation policy has stayed constant: FCFS, OAFA and FREE. Some effort has gone into improving deployment efficiency by spelling out allocation policies explicitly, but the general characteristics of address allocation haven't changed so far. They're still FCFS, OAFA and FREE, no cap 💅.

Here's a possible scenario: pricing processes imposed by the address allocation agency might not be feasible in the current Internet environment. Even if they were, they probably couldn't push deployment efficiency up to the levels needed to extend the unallocated pool's lifetime in the long run. Not being able to use pricing as a management tool, even just to recover the costs of allocation and ongoing registry maintenance, could lead to a few longer term outcomes:

• a) such functions will be restructured and operated from duly authorised national administrative bodies for each nation. Here's the tea ☕: the address pool delegation sequence in the current Internet environment hasn't been lined up with recognised national public communications resource administrative entities so far. That mismatch is a big sign of the core problem, which is that the unallocated pool isn't seen as basically the same kind of public resource as the radio spectrum or the telephone number space. Because of that mismatch between existing public resource management structures and IPv4 address space management, running this activity publicly on a national basis isn't something we commonly see. We really have to question how competent those established public resource management structures are at handling a technology-driven domain that is still incredibly vibrant and dynamic 🤔. A rational and equitable address space management mechanism might come out of it, but in all likelihood it would also come with a heavy damping effect on further technological innovation and refinement of the technology base the address space sits on, and that would show up as a longer term outcome 😭.
• b) such functions are operated (and/or funded) by Internet Service Providers. This is more common right now in the Internet IPv4 environment. Even though this setup does allow for enough funding to run things competently, these entities have established interests in running services across the address space (basically, they're really invested in exploiting it). That can naturally lead to established interests dominating the market, which threatens fair access to the common resource and threatens an open market for deploying the technology. It's fair to say that these kinds of alignments are undesirable from a public policy perspective 🚩.
• c) such functions are inadequately funded for the level of activity, and/or are administered informally and therefore managed poorly. In that case, the essential attribute of reliable address space management just isn't achieved 💀.

It's worth noting that these issues are still largely unresolved within the Internet community today. Tensions between established and incoming Internet Service Providers over equitable access to the unallocated address space pool are a consequent risk 👀.

---

5. Concluding Observations 🎀

Here's the situation: without being able to price management of the Internet address space at administrative cost levels, let alone set address leasing prices that reflect the finite nature of the resource and its market value (even partly), the most likely outcome is that FCFS, OAFA and FREE management policies just keep going until the unallocated pool runs out 🙃. That's the default, bestie.

It's kind of a sad reflection of how short term goals clash with longer term considerations. The short term motivations that made ready and equitable access to IPv4 addresses a priority, which were the very factors that shaped the current Internet address allocation policies, carry a real risk. In the longer term, they could turn into monopoly-based restrictive trade and barrier-based pricing once the unallocated address space runs out 💔.

Free address allocation and policies with pricing components both end up creating strong pressure for better address space utilisation efficiency. ✨ But removing the neutral presence of the unallocated address pool does bring a serious risk of open market failure within the Internet itself, if free address allocation keeps going until the pool is exhausted.

Strengthening the current FCFS, OAFA and FREE address allocation policies, in an effort to get higher utilisation efficiencies across the remaining address space, is not a viable refinement of address management strategy. The reason is that a trading market would start before the unallocated pool runs out, trading large address blocks that those strengthened allocation policies are designed to rule out 🙅‍♀️.

The most negative bit of all of this is that these processes will erode confidence in the Internet community's ability to regulate itself. The wider community is likely to seriously doubt that the Internet process is a good fit for effectively formulating common administrative policy for one of the core shared assets of the Internet 😩.

All of these outcomes can be read as policy failure outcomes. 🥀

How serious these outcomes are has to be judged against the expected timeframe of that policy failure. Current expectations put the unallocated address pool's lifetime at 6 - 12 years. That does give the Internet community some time to revisit how it defines its administrative processes. But that breathing room is tempered by the IPv6 process and by growing pressure on the address space, from rising address demand as the Internet itself keeps growing 📈.

It's probably fair to conclude that the existing processes are an obstacle to admitting any significant process or policy change that would produce a more efficient and effective address space management regime 🧱.

Still, it's this policy failure to efficiently use the IPv4 address space, due to inadequate address pool management policies, rather than the exhaustion of the pool itself, that is perhaps the real driving force behind designing and deploying an evolutionary technology for IPv4. That technology has a significantly larger address space as one of its major attributes 🚀.

It's also fair to conclude that any outside observer of the IPv6 refinement process will look for evidence of experiential learning in address management policies. If a successor technology to IPv4 is going to happen, it would be reasonable to expect that its associated address pool management mechanisms show a better understanding of public resource space management capability, informed by this experience. If no such evidence shows up, there's no clear mechanism to build enough consumer and industry confidence for large scale public deployment, no matter how good the successor technology's technical attributes are 😤. Such mechanisms may include pricing components regardless of the actual size of the address resource, given that the uniqueness of a number is a resource with inherent market value, whether or not scarcity pricing premiums are relevant in that address space 💎.

It's also fair to conclude that continuing with current address space management policies carries a very strong risk of restrictive and monopoly-based trading in address space. The same trading practices would then show up within the deployed Internet itself 🕸️.

The immediate action that seems most aligned with the interests of both the Internet community and the broader public is to examine Internet address space management structures that include pricing as well as policy components within the overall management mechanism. It also means looking at how such mechanisms could apply to the existing IPv4 address space, and to any refinement or successor Internet technology base 🌈.

---

6. References 📚

[1] Gerich, E., "Guidelines for Management of IP Address Space", RFC 1466, Merit Network, Inc., May 1993.

---

7. Security Considerations 🔒

Security issues are not discussed in this memo. 🙈

---

8. Author's Address 💌

Here's where you can find Geoff 🫶:

Geoff Huston
Australian Academic and Research Network
GPO Box 1142
Canberra  ACT  2601
Australia

Phone: +61 6 249 3385
Fax: +61 6 249 1369
EMail: Geoff.Huston@aarnet.edu.au