Diversity of Internet makes IPv6 transition difficult

By : |October 21, 2011 0

[image_library_tag 408/14408, align=”left” title=”” height=”215″ alt=”” hspace=”7″ width=”150″ vspace=”7″ border=”1″ ,default]BANGALORE, INDIA: “In the early period of Internet, we were able to translate from one protocol to another (known as Flag Day), because Internet was a small enough to co-ordinate. However, today it is very distributed in nature. Many governmental and political entities are using it and so there is no way to have a uniform Flag Day, so we need to have other technologies to make this transition happen,” says Truman Boyes, leader of the network architects that comprise Juniper’s Asia Pacific professional services consultancy team.

He was speaking to Deepa Damodaran of CIOL during an interview, recently. Excerpts:

CIOL: Which are the technologies that allow this transition to IPv6?

Truman Boyes: There are several technologies that allow this transition. One such mechanism is Large Scale NAT (LSN), which is also known as Carrier Grade NAT.


It can scale massively to translate one protocol to another and will allow service providers to deploy IPv4, and IPv6 services and thus communicate with clients or other organisation, which are using one or the other protocol.

CIOL: However, reports say that even if you are able to translate to IPv6 with LSN, the network is incapable to scale and support so many users and thus slows down?

Truman: The problem with LSN is that it eventually becomes a part of the NAC in network. So while selecting the the scale of a product, a company should apply a lot of diligence and select the right product.

The type of scale that we are talking about is close to 100 million sessions through a single box.

So if something happens to that particular box or device on the network, then it will have a huge effect on millions of computers at one point of time. So a lot of rigour has to be put into in the types of products that fill the role.

When you are doing translation of protocol, some application do not respond well with the process.

There is less and less of these kind of applications because almost everyone has a router at their home, which is capable of performing NAT (Network Address Translation), although at a smaller scale. So most applications that we use today work well with NAT.

However, there are a few business applications that do not, and if such products are there in the middle of the network, which is being used for translation, it will create issues for this application.

CIOL: So what should be done to applications so that transition to next generation protocol is smooth in the future? Why do companies hesitate to move on to IPv6?

Truman: In case of Carrier Grade NAT, one thing that needs to be changed is the way in which application designers write software.

All application designers need to move away from embedding addresses into application or code because when we are translating a particular part of the Internet packet from IPv4 to IPv6, or the other way round, and if there is something that can not be translated and if it is sitting in the middle of the packet, then the application will not to work properly.

Moreover, in terms of adoption of IPv6 there is a complete lack of strategy. It is more of an organisation issue, than a technical one. A lot of organisations do not have an understanding of how they are going to move to IPv6.


The second important thing is the cost to replace legacy equipments. The cost of routers, that sit in a fixed line network, is very cheap at less that 50 to 100 US dollars. However, if you think of replacing all old equipments and translate them to IPv6, it will be very expensive for consumers.

Deploying IPv6 has become more of a chicken vs egg scenario today.

Internet providers or content providers are not deploying IPv6 because there was no IPv6 content. It was not being delivered to consumers because there were no consumers who natively have IPv6 on their device. So far there was no real push for content providers to produce IPv6.

CIOL: What is all this discussion about ‘Two Internet’, one for IPv4 and one for IPv6?

Truman: No matter whether it is IPv4 or IPv6, we will still have one Internet
Some servers, some content providers will provide service to IPv4 clients and IPv6 clients and effectively they bridge the services together.

Today Google is available in both the versions and it is completely the same service on both. Though the Internet will be one still there will be some concerns regarding disparity in the two or whether they will be equal or not in terms of performance.

Technically, they should be the same, however since IPv6 is not yet deployed by all service providers across the world, service traffic will not route through them exclusively. So there might be a variation in performance between IPv4 and IPv6, which will probably change overtime.

CIOL: So what will ultimately happen to IPv4?

Truman: We expect IPv4 to be around for a very long time. It could be ten or fifteen years. It will be very long time before IPv4 is completely exhausted.

Even if we move all the services to IPv6, which will happen in the next 10 years, there will be several devices on the Internet that use IPv4. There are so many devices or hardware that will not be upgraded because there is no reason to upgrade them until it fails. So IPv4 will be around until those devices are no longer technologically relevant and are de-commissioned.

So we will still live with IPv4 and hopefully not even know the difference between the two.