The world is running out of IPv4 address space to allocate. Whilst some could be reallocated from the original large blocks issued in the early days, this could cause problems in the size of routing tables in the core internet routers. Will this lead to an opportunity to replace domestic customer equipment with IPv6 capable products that could include femtocell technology too?

Does this sound like the end-of-the-world in 2000 scenario again?

At the end of the last millenium, forecasters of doom were advising businesses that all their software should be checked to ensure it would continue working after 1999, leading to a healthy workload for software businesses. Whether you believe the resulting lack of disruption was due to the heroic efforts to validate and/or replace software systems before the deadline, or that there wasn’t really a big problem to avoid in the first place, it certainly caught people’s attention and attracted heavy investment.

The clock is ticking – time is running out

A similar story is being promoted by those looking after the core internet address space. The Number Resource Organisation warns that over 90% of IPv4 address space has been allocated, as reported by ARS Technica. SmartPlanet suggests there are less than 700 days before exhaustion, attracting a variety of comments both supportive and cynical.

You may have heard this issue cropping up before, with a variety of workarounds having been created to articificially extend the address space. In particular, using Network Address Translation (NAT) to multiplex large numbers of devices to a single IP address using port numbers has been very effective. But so far, the internet continues to grow with more users, traffic and services being added daily – and still seems to work. Many may believe there are still workarounds yet to be brought into use.

Whilst IPv6 originally touted a wide range of great new features and capabilities, most of these have been ported back to IPv4 (such as Secure Sessions, Quality of Service). The main reason why any operator would make the transition is to solve problems with address space, rather than to provide new capabilities. It can also be argued that by giving each individual device their own unique address, rather than sharing/dynamically allocating one, then specific management functions and peer to peer applications would operate better. I’d say these aren’t “show stoppers”, just additional good reasons to encourage the migration.

Why would businesses, customers and networks migrate to IPv6 anyway

The commercial driver to adopt IPv6 seems to be coming down to that of address space alone. Some, like Comcast – the largest cable operator in the US – are well advanced in their deployment and are said to be gearing up for mass residential deployment in 2010. Others offer this as an option (my own ISP for example provides me with an IPv6 address at no extra cost).

But there seems little immediate benefit or driver for me to take this up. Without looking into it in great detail, it seems to me that my web browser would work more slowly if based in IPv6 while it tries everything with the new protocol first before then falling back to IPv4.

The benefit is for the whole community (myself included) and needs to be driven by network providers rather than consumers.

LTE mandates IPv6

In between the blaze of headline data rates and high performance specifications of LTE, the new 4G radio network technology, is the requirement to connect cellsites through IPv6. This will also encourage wider takeup and availability of the technology. 3gamericas has published an excellent white paper on the topic. It explains that dual-stack (IPv4 and IPv6) may be likely for some time.

Standards already allow end user mobile devices to choose between IPv4 and IPv6 on 3G, although I doubt if there is much usage of it so far. Desktop PCs also have some of this technology built in and ready for use in due course.

This implicitly means that LTE femtocells would need to be IPv6 capable and connected back to the core network using this technology.

Some other drivers for IPv6

Other reasons for network providers to make the investment include:

1. Greater difficulty for core internet routers to handle the fragmented IPv4 address space. As the space gets further divided up and allocated, routing tables will become much bigger and take longer to lookup and process each packet.
2. The growth in peer to peer traffic applications (not illegal file sharing) would benefit from using IPv6, where each device can have its own static address.
3. Stock market valuations for network operators may be affected by the presence (or otherwise) of their IPv6 migration plans. It is said that some stock market analysts have checked to see if these exist and are pragmatic.

The likely path to adoption

This won’t be a sudden switchover. We can expect to see more and more ISPs support IPv6 as well as IPv4 in their networks, a so-called dual-stack approach.

The core internet will develop to support IPv6 as quickly and effectively as IPv4 today.

New end user equipment will become to be IPv6 capable and use it where possible. Some equpiment is already capable of this today.

Websites will upgrade to be IPv6/IPv4 dual mode and advertise this in their DNS entries. In most cases, Website hosting companies will deal with this complexity and it won’t affect the content, format or operation of the sites.

The reason why I believe websites will be last is that there could be long delays when trying to access sites via IPv6 where some link in the network doesn’t support the protocol.

Unfortunately this reduces any incentive for consumers to make the change.

Will this drive set-top boxes/modems to be replaced by IPv6 capable devices sooner rather than later?

Where operators do choose to under a major upgrade to their customers domestic internet modem/routers over the next few years to meet this challenge for IPv6 deployment, wouldn’t it make sense to incorporate femtocells in these boxes and use that as an additional benefit for the customer to upgrade?

With significant investment required in the next few years for both technologies, wouldn't this make commercial sense?

And finally

Apologies if you think this is off-topic for femtocells. The data crunch hitting the telecoms industry is coming from several fronts, not just capacity, and forewarned is forearmed.

PS: Even more off-topic, just in case you ever wondered what happened to IPv5, this was an experimental extension to IPv4 which handled connection oriented streaming traffic. Many of these concepts were incorporated into MPLS instead.

Articles for further reading

http://arstechnica.com/articles/paedia/IPv6.ars/4

http://www.potaroo.net/ispcol/2008-10/v4depletion.html

SmartPlanet - 700 days until IPv4 addresses run out

Comments (0)