IPv6—It's Not Just About Software Anymore!

Dec. 8, 2004
Discussions of the migration from Internet Protocol version 4 (IPv4) to version 6 (IPv6) have been lively and sometimes contradictory. Opinions vary widely as to why it is needed and what role stopgap measures, such as network address translation...

Discussions of the migration from Internet Protocol version 4 (IPv4) to version 6 (IPv6) have been lively and sometimes contradictory. Opinions vary widely as to why it is needed and what role stopgap measures, such as network address translation (NAT), will play along the way. In spite of the many apparent contradictions, the question is not if the transition is coming, but rather when we can expect it.

Momentum is growing for a transition to IPv6 as millions of new users deplete available IP addresses under IPv4 and demand grows for new services and increased security. Despite this momentum, the deployment speed of IPv6 unexpectedly varies from one geographic region to another. Government-driven efforts are well under way in several parts of Asia, where the Internet address space is already constrained by the rapid growth in new mobile phone users. At the same time, European governments are also laying a foundation for an eventual transition to IPv6.

The U.S. has experienced a less urgent need for the transition due to large IP address allocations and the use of NAT. However, NAT's ability to conserve address space can affect the rollout of compelling new applications that rely upon a peer-to-peer model, such as VoIP and file sharing. As the private sector debates the effectiveness of NAT and the deployment of new services, the Department of Defense has announced plans to transition all of its inter- and intra-networking to IPv6 by 2008. Is it just a matter of time before commercial markets follow?

Much of the technology debate so far has focused on the software implications of the migration. What impact will IPv6 make on hardware architectures? Given the long design cycles of network equipment and the time that networking equipment is deployed in the network, doesn't it make sense to consider the implications now? Waiting to consider the hardware issues of IPv6 until a customer asks for it or the government mandates it is a risky strategy.

The most notable change in IPv4 to IPv6 is the increase of address sizes from 32 to 128 bits. Under IPv6, a typical policy lookup is 296 bits or longer, including an 8-bit IP protocol field, the 128-bit IP source address, the 128-bit destination address, the 16-bit TCP source port number, and the 16-bit destination port number. The increase in address size will place heavier demands on equipment forwarding and policy evaluation performance. Policy lookups—used to implement functions such as forwarding, quality of service (QoS), filtering rules for ACLs, and billing—will present a much more imposing challenge.

To address IPv6 challenges, network equipment manufacturers will need to build more powerful packet-processing capabilities into their products. However, due to the uncertainty of the IPv4 to IPv6 transition timeframe, network equipment will also have to retain backward compatibility with IPv4 while efficiently supporting the larger header sizes used in IPv6. Packet-processing ASSPs such as network search engines (NSEs), with their wide search widths and high performance, will be critical to delivering optimized network equipment architectures that require 200 million searches per second (MSPS) and beyond.

Beyond the performance and functionality requirements to enable the IPv4 to IPv6 transition, an equally important need exists for a new level of reliability to support new-generation services that will roll out during this transition time. More stringent service level agreements (SLAs) will place new demands on networking equipment that translate into new requirements for all critical components in the architecture. Packet-processing solutions must specifically address reliability issues such as soft errors to support the "five nines" availability that major service providers need to fulfill their service commitments.

It is still anyone's guess when the transition from IPv4 to IPv6 will hit critical mass. But that doesn't mean hardware designers should wait for widescale deployment to begin. The time to examine the implications of IPv6 on your architecture and how you can future-proof your design is now.

Sponsored Recommendations

Comments

To join the conversation, and become an exclusive member of Electronic Design, create an account today!