The Policy Differences Between IPv4 and IPv6
Policies are designed to achieve goals. Both IPv4 and IPv6 share four common policy goals. They are designed to ensure that internet users have IP addresses, can use them, and network operators can coordinate with each other when there are technical problems. The four common goals are:
- Uniqueness: avoid conflicting claims to a public IP address
- Aggregation: hierarchical distribution means routing information can be summarized
- Conservation: fair distribution to end user networks
- Registration: a public registry to aid troubleshooting
But IPv6 has an extra goal and a decider. It’s worth looking at them to understand the different challenges these two versions of the Internet Protocol bring to organizations and the engineers running networks.
An Extra Goal for IPv6
Those common goals were formulated for IPv4. And IPv4 – a 32-bit address space – is pretty small. It has just 4.3 billion addresses, fewer than one for each person on Earth.
IPv6 is much bigger. It has a 128-bit address space, giving 340 undecillion addresses. People struggle to understand such a number, much less the concept of addresses available in that quantity. But those extra addresses bring problems as well as solutions, so the policy goals need to be expanded.
A new goal is minimized overhead, which means fewer requests for more addresses. But aggregation is held out as particularly important in IPv6 policy because of the potential impact on routing.
Aggregation, or Hierarchical Distribution
Networks share information about how to get from one place on the internet to another. They use the Border Gateway Protocol, or BGP, to share these internet routes. Networks share their view of the addresses they know about and routes available to get to them. In early 2025 there are about 220,000 IPv6 routes on the internet and just over 1 million IPv4 routes.
On the internet, a route is the path taken between two networks. Networks announce the blocks of IP addresses they have to the networks they directly connect with. Those announcements are passed on the other networks. But a block of IP addresses can be announced as a whole – the aggregate – or broken up into pieces, called “more specifics.” About half of all BGP announcements are more specific announcements. This is often done to get traffic to take particular routes and is called traffic engineering.
But IPv6 is vastly larger than IPv4. That means disaggregation – announcing a block of IP addresses in multiple parts – could lead to a much larger number of routes. And every route adds cost to other networks.
As networks change their connections to each other, connected networks must recalculate their view of the internet. Some will lose access to a network prefix, others will see it through a different network connection. More routes means more recalculations. If the number grows faster, the replacement cycle for network equipment has to speed up to use faster processors. And that can make internet access more expensive for everyone.
So, encouraging the allocation of fewer, larger blocks of IPv6 address space is advantageous in reducing the number of internet routes. This keeps costs down.
This goal is marked out as being particularly important in all five RIRs’ policy documents.
Minimized Overhead, or Administrative Ease
Internet engineers realized that IPv4 address space was running out in 1992. By 1996, the expectation was that the registry would allocate a block of addresses that would be half used in a year. You could come back for more when you had just a fifth left. But IPv6 is different. There’s no shortage of addresses, so there is a policy goal of reducing the number of times a network operator will need to return to the pool.
Each time an organization needs more addresses they must prepare an application, which must be reviewed by the registry. The RIR must take a decision on whether to allocate addresses, and if so, how many. That’s a burden on everyone and pushes up costs.
IPv4 policies had a “slow start” built in to avoid the risk of allocating a large block of addresses that would never be used. That risk does not exist in IPv6. Despite about half of all internet traffic using IPv6 and some huge allocations, less than two percent of the IPv6 space has been allocated. So, the RIR communities – engineers, regulators, and business representatives – decided against a slow start mechanism for IPv6 allocations.
Implementation
A single policy choice can support both these goals.
That choice was a generous initial allocation. It is large enough to meet most organizations’ foreseeable needs. That means no complex assessment of the organization’s network plans.
It also means that organizations will always be able to announce a single block of addresses, instead of a collection of different blocks. They might cut up their block sometimes. But if the need for that goes away, it can be put back together again.
The RIRs make generous minimum allocations. They all allocate at least a /32 without assessing the requestor’s needs. The RIPE NCC allocates up to a /29 on request. A /32 is enough for up to 2 million residential customers and a /29 is eight times larger.
What’s more, they all allocate much larger blocks when a need is demonstrated. For instance, ARIN has allocated a /16 to a large finance company. That’s 65,536 times larger than the minimum. And more than 50 organizations have got /24s, that’s 256 times the minimum.
Managing IPv4 and IPv6
Whether you run IPv6 or IPv4, you’ll need to manage your addresses. ReView is a free tool from IPv4.Global working with 6connect that can do just that. It can help you inventory your addresses, so you can be sure you are in control of your assets.
If you need IPv4 addresses, you’ll probably need to go to the market. IPv4.Global’s auction platform is the most transparent marketplace available. We can help you find the addresses you need, and we can connect you with experts who can make your addresses meet your business needs.