Monthly Archive for March, 2011

This is the first of two posts on IPv6 written by Jag Bains, Director of Network Operations at PEER 1 Hosting. You can read his second post entitled Current IPv6 Network Techniques and Transition Strategies.

Over the last few months, staff at PEER 1 Hosting have been coming to me and my network colleagues to understand what all the IPv6 racket was about. They were wondering if the Internet as we know it was going to crumble now that IPocalypse has come and gone. To these concerned I replied, “Time to renew your favourite magazine subscriptions, because the Internet as we know it is dead”.

Nanobots Cartoon Credit: http://xkcd.com/865/

All kidding aside, there has been a lot of confusion in trying to understand where this sudden momentum for IPv6 has come from.  There have been numerous articles, even within mainstream media, that discuss how IPv4 has been exhausted. Cautionary tales and tech sermons from the silicon mount are being delivered advising enterprises, service providers, governments–basically anyone with an Internet presence and requirement–on how they need to start incorporating IPv6 strategies into their operations. But what does that really mean? To better understand why the need for IPv6, let’s take a very quick look at world of IPv4 and the challenges it faces today.

Let’s kick it off at a very rudimentary level. The Internet is basically a collection of networks, interconnected and identified to each other by their unique Autonomous System Number (PEER 1 Hosting’s ASN is 13768). Each ASN has at least one Class C, or 255 contiguous IPv4 addresses, in order to be routable on the global Internet. PEER 1 Hosting has well over 1.5 million IPv4 addresses. These IPv4 addresses are provided to service providers such as PEER 1 Hosting by one of the five Regional Internet Registries (RIRs), which are organizations that work together to provide technical co-ordination and management of IP addresses.

The RIRs receive their IP addresses from the Internet Assigned Numbers Authority (IANA), which is the entity that oversees, among other things, global IP address allocation. If you want to see a great technical breakdown of the relationship between the RIR’s, IANA and every IPv4 address, feel free to check out this article.

So now we understand how an IP assignment from IANA makes its way down from the local RIR, to the Internet Service Provider (ISP) network, and ultimately to the end user’s desktop or server. But the big question is, “What happens now that there are no more IPv4 addresses to allocate?”

IPv6 to the Rescue… Sort Of
In early February 2011, the pool of unassigned IPv4 addresses was depleted when IANA delegated the last five blocks of IPv4 address space to the five RIR’s. The RIR’s are expected to dole out the majority of these IPv4 addresses to carriers before the end of 2011. What the rate of IPv4 exhaustion will be within each carrier will vary from one entity to the next. Some models like mobile broadband are running out faster than others, but it is safe to say that most companies won’t have a surplus to last them any significant time.

Fortunately, the Internet Engineering Task Force (IETF) were prescient enough to know that the 4.3 billion IPv4 addresses were not going to be enough over time, and developed the RFC for IPv6 back in 1996 to address this shortage. The RFC called for a new numbering scheme that allowed for 2¹²⁸ addresses (as opposed to 2³² ). With over 340 undecillion IP addresses in IPv6 (340,282,366,920,938,463,463,374,607,431,768,211,456 for those of you that were curious), IP resource exhaustion is not going to be a concern anytime in the foreseeable future as it was for IPv4.

Ipv6 is so large, in fact, that hexadecimal is needed to represent the size of allocations within IPv6. This larger address schema not only allows for many more devices and users on the Internet, but it also provides extra flexibility in allocating addresses and efficiency for routing traffic in comparison to IPv4. The designs for IPv6 also brought in additional capability in comparison to IPv4, notably:

• Multicast – IPv6 forgoes on the idea of a traditional broadcast address and uses a link local address and rendezvous point addresses, allowing for great flexibility in creating inter-domain multicast groups. This has huge potential for video streaming strategies.

• Stateless Address AutoConfiguration (SLAAC) – IPv6 hosts can auto configure themselves in a routed IPv6 network using ICMPv6. This can be very useful to large cable/DSL network that have multiple hosts constantly being added and removed from their networks

• Embedded support for network security – IPSec, the authentication and encryption of each IP packet, has been back engineered in IPv4, but it was originally developed in the designs for IPv6.

• Packet design – IPv6 redesigned the way the headers (similar to your mailing address with special instructions) are built to make the addition of custom headers/instructions efficient and scalable.

Using the same model of delegation as IPv4, organizations have been procuring IPv6 address space and formulating game plans on how to implement into their networks. Staying in step with the network world, technology manufacturers have also been incorporating IPv6 capabilities into their server operating systems, firewalls, load balancers, and other various Internet products for the last 5-8 years. It’s safe to say that most network and server farm environments now have enough capability to run IPv6 services. The question now is: how are they going to implement it?