250 million-plus reserved IPv4 addresses could be released – but the internet isn’t built to use them
A new chapter in the long saga of the 240/4 block is being written. If you want more and cheaper IPv4, maybe you should help
Comment Activists are again lobbying for more than 250 million unused IPv4 addresses to be released for use, potentially tackling the IPv4 exhaustion problem. However, the proposal has been tried and failed before, and again faces formidable opposition.
The unused addresses are known as the “240/4” block and comprise most of the IPv4 addresses from 240.0.0.0 to 255.255.255.254 – a space that encompasses upwards of 268 million addresses, or about six percent of the entire IPv4 number space. For some perspective on the magnitude, at market rates, the addresses are worth around $7 billion
All this potential sits unused because in the early days of IPv4 they were set aside for future use or experiments.
Those early years of internetworking were a bit like the universe in the aftermath of the Big Bang: things happened very quickly, under strange conditions, with some scattered oddities surviving. We still see some of those survivors (Pi and Pythagoras) along with a fair bit of detritus.
Decisions made in those early years of the internet have halted a fair bit of evolution. Consider, for example, that 16 million IP addresses were allocated to ham radio operators, and the protection of 240/4 for future use.
Decades later, both decisions appear anachronistic, not least because the ham radio operators recently made more than $100 million by selling some of their IPv4 assets to Amazon – an extraordinary windfall and one that rankles given that the world all-but-exhausted its supply of fresh IPv4 addresses over a decade ago.
IPv4 exhaustion has seen prices for a single address reach $30 or more on the private market, a handful of dollars each a year when leased from specialist providers, or $43.80 a year when rented from AWS.
The shortage of IPv4 and the cost of acquiring new resources has sparked many to ponder whether it’s time to re-classify the 240/4 block and even allow it to be administered by regional internet registries that would allocate them to members at modest cost.
One person who has tried to change the status of 240/4 is Paul Wilson, director-general of regional internet registry (RIR) the Asia-Pacific Network Information Centre (APNIC).
In a 2008 IETF draft Wilson and his co-authors proposed re-classifying 240/4 so it could be used for private networks. There are already some IP ranges reserved for that purpose, namely 10/8, 172.16/12, and 192.168/16. In the 2008 draft, Wilson and colleagues suggested 240/4 could be added to those pools to help “large private Internets that require more address space than is available in the private use address space designated by [RFC1918] during the dual stack transition to IPv6.”
Wilson told The Register that opposition to the idea of changing 240/4’s status came from those who felt it would take a decade or more to implement.
The reasoning is that many manufacturers of networking equipment don’t recognize 240/4 and simply won’t process packets sent to the millions of addresses the block contains. Doing so makes sense: why confuse users by letting them access addresses that aren’t connected to the public internet?
Even if a policy change freed 240/4, it could be years before it was practical for use - especially on public networks. To understand why, consider the billions of home routers, most of which won’t recognize 240/4 or would need new firmware to do so. If the block were freed for public use, some of its IP addresses could therefore be invisible to many users. Workarounds are possible, but many internetworking wonks are already a bit ashamed at the prevalence of network address translation!
- ICANN proposes creating .INTERNAL domain to do the same job as 192.168.x.x
- China requires any new domestic Wi-Fi kit to support IPv6 and run it by default
- APNIC close to completing delegation of its final /8 IPv4 block
- AWS: IPv4 addresses cost too much, so you’re going to pay
Those issues didn’t stop others from trying again to free 240/4. Some folks from Cisco had a crack in 2008.
Also in 2008, a Linux patch saw the open source kernel recognize the 240/4 block. The Linux-derived router firmware project OpenWRT did likewise in the same year. MacOS and Solaris have also recognized it for years. OpenBSD started doing so in 2022.
Some versions of Cisco’s IOS also support the block. Juniper’s JUNOS can be made to at users’ discretion.
Big tech's 240/4 backdoor
The topic of 240/4 has been raised many times over the years, but Big Tech has drawn fresh attention to the block by using it internally.
The authors of the piece at that last link are members of a group – The IPv4 Unicast Extensions Project – that is now advocating for a change of status to 240/4.
The group imagines slow change to ensure that the block doesn’t disrupt the operation of the wider internet. The Project is pursuing its goal through the standards process: here is its IETF draft.
One of the Project’s members, Seth David Schoen, told The Register that opposition to the plan has come from those who use the old argument about the need to replace or upgrade potentially billions of devices to ensure universal access to 240/4, citing the cost, risks, and potential ecological impact of upgrading the world’s networking device fleet. Others have argued against it on grounds that freeing IPv4 addresses will further retard migration to IPv6, a protocol with a practically infinite number space and many fine features that IPv4 lacks.
Interop testing needed
Schoen said the project hopes to test networking devices to better understand if they’re able to be upgraded to handle 240/4. He feels most could handle the change as very little code is required to change the values of IP addresses a device will access.
The giant switches and routers used in the internet's core routers may be more of a challenge: Schoen said Project members understand that some may have the IP ranges they access set in silicon, the sort of thing such high end machines do to improve speed. If core infrastructure can’t easily be made to access 240/4, it would be a significant barrier to its release.
Australian network engineers Karl Kloppenborg and Christopher Hawker are also calling for change to the status of 240/4 due to IPv4 exhaustion. The pair point out that APNIC currently allocates those who apply for IPv4 addresses a /23 block - just 512 IPv4 addresses - when it has some to hand out (usually when unused addresses are returned to the registry for re-use). Kloppenborg and Hawker believe that if APNIC were given just three of the 16 /8 blocks from 240/4, the registry could satisfy demand for new IPv4 allocations for 20 years or more under current allocation policies.
The two feel that, in the hands of RIRs, the 240/4 block could therefore allow developing nations where telcos may not be ready for IPv6, and skills and equipment are scarce, to enjoy the benefits of greater competition as local companies use IPv4 resources to offer services.
Kloppenborg and Hawker plan to propose APNIC consider getting behind a change of status for 240/4 at the APRICOT conference later this month.
Schoen and his fellow Project members are trying to arrange tests, and continuing to seek support for their draft.
All of which leaves 240/4 in the hands of community processes, and therefore in the hands of community members, even as competition for IPv4 resources becomes fierce and those with deep pockets – like AWS – monetize a resource that was once handed out without thought for its future use.
240/4 reform offers a chance to revisit such decisions and to influence the fate of IPv4 itself.
That the old protocol will be with us for decades to come is not in doubt.
But 240/4 reform could, perhaps, stretch or shrink the number of decades in which IPv4 remains relevant. ®