Rick van Rein

zo 16 juli 2017


NGI 1: IPv4 is Killing Progress

IPv4 is no longer possible without NAT, which holds back the development of the Internet. In short, IPv4 has served its time and should be forcefully replaced by IPv6 in the coming 5-6 years.

This is part of a series of technical articles in response to the European Commission's initiative to explore the Next Generation Internet, an initiative that we warm-heartedly support.

Among the many concerns that protocol innovation faces is the reality of having to deal with NAT routers. This is fundamentally more difficult than dealing with a firewall, in a way that blocks progress.

A good example of this is the ability of peers to connect directly. Crossing through firewalls is easy enough when the peers intend to collaborate, but the added complexity of a different public address/port from what is used internally, and specifically the inability to know for certain how things are mapped make it impossible to make such peer connections.

As a result of this, we see many applications fail. Passing a document over to a collegue on another network, for example. Or simply conducting a phone call or chat session directly over the Internet. Other application protocols are conceptualised but fail in the design phase, on grounds of "shifty" remote addresses/ports. All these things are blocked by the NAT concept, and therefore, by the over-extended life of IPv4.

One of the direct results is that all protocols in use today are client-server protocols. Meaning, someone with better technical equipment than an end-user places a server online, and all communication must pass through their systems. This cannot be done without gaining something in return; too often, this ends up in perverted privacy. End users have no choice but to live with that. Had protocols been more able to run directly between peers, as is the case when IPv6 can be counted on, then this assault to one's privacy could be resolved by more direct communication patterns, including peer-to-peer.

As soon as IPv6 is introduced, these clouded skies clear up. No more NAT means transparent addressing and no more "shifty" remote addresses/ports. The freedom that this brings to protocol designers is incredible, and true innovation can be expected once we get on with IPv6.

But IPv6 is not sufficiently sexy to catch the public attention. That is because the change is not noticable at the user level; and that is caused by a catch-22 situation where applications don't count on IPv6 because it isn't there, and it isn't there because applications don't count on it. Also, IPv6 only works when it is available on two sides, so one-sided investment in the extra work barely pays off; certainly not in general. All this holds back the development of new protocols that are vital for a healthy Internet, where users are better able to sever the bonds of online service providers. In short, the only way for this long-overdue update to the Internet if vital parties like the European governments stand up and start requiring it from anyone who deals with them.

It is possible to have IPv6 anywhere. It is a matter of getting enough critical mass, of making it a necessity on the Internet. Then, options exist, including the work we've done in the Peer-to-Peer Toolkit and for which an SIDNfonds innovation budget was awarded. Specifically, the 6bed4 tunnel is a European design that does what Microsoft's Teredo wanted to do but at which it failed utterly: to enable IPv6 on every machine on the planet. Unlike Teredo connections, 6bed4 can be relied upon to connect IPv6 everywhere.

The time is here to activate the move away from IPv4, and the road is simply making it a requirement. My proposal is the following:

  • All new ICT deliveries to the governments of Europe and its member states must support IPv6; products failing to do so will be phased out over a period of five years;
  • In the year following these five years, all European governments completely shut down their use of IPv4. Without remorse. This means becoming IPv6-only, through removal of backward compatibility with IPv4 from public-facing servers, of course including for web and email.

Such bold action, requiring all others to follow suit, forces the rest of the market to follow. This appears to be the only way to bootstrap the long-overdue transition to IPv6, and reduce the protocol headaches of IPv4 mere remnants of the past. Citizens throughout Europe, and probably far beyond, will recognise the need to facilitate IPv6 to be compatible with European governments; technical facilitators will find their customers demanding it. Protocol developers can now focus on an Internet that enables new protocols and once more experience a functional Internet. This way, Europe quite literally leads the way towards the Next Generation Internet!

The period of 5 years is based on the time it takes to replace devices. Device manufacturers and Internet Providers have long known that they should prepare for IPv6. Nevertheless, homes and offices may now still contain devices such as printers that need to be updated, for which 5 years is a reasonable and pragmatic time (because devices rarely live longer). It is even more than reasonable, as the message to upgrade to IPv6-capabilities has been sent for many years already.

The only thing needed now is a switch, which the free market clearly is not organising on its own. And this is where an important, large enough usage base should take the lead. Who should do that? It is not likely to be commercial businesses. The only party truly empowered here is the collective government infrastructure of Europe and its member states.

I expect an outcry from some Internet Providers about this -- but there is no need, there are many ways in which they can change over in the coming 5 years, it has all been shown to work by the few pioneers. Mobile networks too have developed the technology, it simply needs to be switched on. But outcries will come, and they will originate from precisely those Internet Providers that have been ignorant of this innovation that has been hoping for free-choice adoption for years; this freedom to hold back the Internet will indeed be taken away, because it is impairing the health of the Internet. Complainers are a sign of the stagnation that landed us with NAT, and they should be frowned upon, rather than supported in their stifling lack of action.

Go Top