IPv6 - The solution for the Next Generation Internet?

Introduction
Today, the Internet is at a point of the next threshold. With the convergence of voice, video & data the existing communication backbones are being replaced by high-speed optical IP backbones. While this is just happening, the 3GPP (3rd Generation Partnership Project) has proclaimed that they have decided to adopt IPv6 as the protocol for future IP multimedia services (May 26, 2000). For the first time now, wireless carriers can pressurize commercial backbone operators to implement IPv6. Until now, IPv6 deployment has been done mostly by nonprofit and scientific organizations.

In this article, we take a look at what is IPv6, the problems it is trying to address, roadblocks and probable evolution scenarios.

What is IPv6?
IPv6 is the next version of the very popular IP Protocol. The current version of the IP protocol in use is IPv4 (IP Version 4). The primary difference is the increased size of the IP Address from 32 bits to 128 bits, allowing more addressable nodes on the Internet!!

What are the problems with IPv4?
The most important problem with IPv4 today is the Acute IP address Shortage due to increasing number of devices (computers, mobile devices, home appliances etc.) connected to the Internet. IPv4 addresses are allocated from a 32-bit address space. Further the addresses are divided as Class A, B, C & D. This has given rise to inefficient usage and wastage of addresses (Class A, B). In addition, unused addresses were not re-used or re-claimed.

The panacea to this problem now is known as Network Address Translation (NAT). NAT allows one to take a single Internet address and multiplex it among various devices, by substituting addresses of packets at the edge of the network (NAT enabled router). However, since NAT peeps into every packet, it is inefficient in terms of performance and creates security/integrity holes.

There is a rapid growth of the applications/protocols that need IP addresses globally scoped, unique and routable e.g. VoIP, videoconferencing (RTP/RTCP protocols), IPSec, Kerberos, games. NAT cannot help here - how does one reach a VOIP application, which has a private IP address?

The technical solution for the above problem (not impossible) will be clumsy and definitely not scalable.

The shortage of addresses is not so acute in North America as it is in other parts of the world, particularly developing countries like China, parts of Asia, Africa and European countries. In fact an ISP in Sweden is only able to hand out private addresses and use NAT to multiplex a few Internet addresses.

IPv4 was designed in the early 1970's. Several new protocols were designed over time, to solve problems as they came up. E.g. Mobile IP - to solve the problem of hand-over with a wireless device or IPSec to solve security related problems. Introduction and inter-working of new protocols requires continuous effort from the ISP, vendor and the users. The proponents of IPv6 feel that it is time to consolidate all the efforts and move ahead with a new protocol, while it is still not too late.

So what else does IPv6 offer?
IPv6 helps network architects revisit fundamental IP issues such as IP address shortage, security, quality of service, mobility, multicasting, and network management. For some of the above, extension protocols have been defined in IPv4 also, but they are optionally implemented. The following are inbuilt into the basic IPv6 protocol as mandatory requirements.

	Increased address size supports more levels of addressing hierarchy (site local, link local, global), greater number of addressable nodes and simple auto-configuration. The addresses have 2 parts - separating the "where am I"(routing prefix) from "who am I"(interface address) part.
	Routing enhancements like provider selection based on policy, cost, auto-readdressing, host mobility etc
	Scalable multicast and "anycast" support
	Header Format Simplification - Fields dropped, fixed size, twice the size of IPv4 header, better handling of options
	Security via IPSec is inbuilt in the protocol
	Support for authentication, data integrity and confidentiality
	Designed to support Mobility, neighbor discovery, auto-configuration
	QOS support - labeling of packets, prioritization and special handling of packet

Status
Let us look at the status of standards, forums supporting this and the implementation status.

Standards: The ipv6forum, which is part of the IETF has come out with standards for almost all protocols (e.g. IP, BGP, RIP, DNS, OSPF etc) and changes required to existing protocols (e.g. UDP). The ipv6forum includes very distinguished companies like Cisco, Nortel, Sun, Ericsson, Nokia, Motorola, AT&T, Sprint and others. The forum has been very active this year, spreading awareness about IPv6 all over the world (next conference in Australia - Sep 2000).

Supporting Organizations / Forums: The 3GPP believes that the success of next generation wireless technologies like GPRS, UMTS, other 3G products and services depend dramatically on the deployment of IPv6. Hence they are strongly supporting it. Realizing that the IPv6 needs a worldwide initiative, several governments are funding projects to further the cause. E.g. 6REN/6TAP (USA), Wide Project (Japan), 6INIT(European Union, Canada), P702/P1009 (Eurescom). These projects are in advanced stages of deployment of IPv6 networks and backbones.

Implementation Status: Deployment of IPv6 may affect most parts of the network. They can be classified as:

Service Providers
ISP's are currently worst hit by shortage of addresses. But, looking at the complexity and volume of changes to support IPv6, they are not all enthusiastic. In fact they are strongly resisting it. Eventually, they may have to yield to the demands for IPv6 related services, if the commercial demand is very high. A few ISP's like Sprint, 6bone are already experimenting with IPv6 services in USA/Europe.

Operating System
OS vendors like SUN and recently Microsoft are upbeat about this. Linux, FreeBSD were among the early proponents of IPv6.

Routers
Most of the major vendors have a Beta implementation of IPv6 in their routers already. Cisco has even upgraded its IOS to support IPv6 in March this year. However support for the entire range of routing protocols and interoperability with IPv4 networks has to be proved.

Networking Products, Tools, and Applications
This is one area, where nothing much has been done. Vendors are waiting for IPv6 to become a reality, before venturing in this area.

Transition Mechanisms from IPv4
One of the most important design goals of IPv6 was backward compatibility and a smooth transition mechanism. Various mechanisms have been specified for IPv6 hosts and routers: Dual Stack approach (complete support for both IPv4 and IPv6 stacks), IPv6-over-IPv4 tunneling, encapsulating IPv6 packets within IPv4 headers to carry them over IPv4 routing infrastructures, IPv4 multicast tunneling (6-over-4) methods etc. The methods to choose will vary from one installation to another, does the ISP support IPv6, what kinds of applications are there and so on.

IPV6 Roadblocks: Another Y2K?
In spite of the technical feasibility and goodness of IPv6, there is enormous resistance to accepting IPv6. Opponents argue that this effort may prove many times the Y2K effort and the amount of destruction it can cause is humongous. Proponents believe that they have defined excellent migration strategies and the evolution can be taken up in a phased manner.

It is unlikely that E-commerce and other high level applications will have to be re-engineered, because of the deployment of IPv6. . However, all IS departments, will have to handle IPv6 clients/servers/routers/applications etc. Hence they must plan and budget for this change.

In spite of best intentions, past experience has taught that any change in a basic protocol is a tedious effort in terms of user acceptance, implementation, testing and rollout. Further, some parts of the specifications are yet to be polished. Demonstrating success through a large enough test-bed network is an issue, yet again leading to lack of confidence.

Conclusion
IPv6 will arrive, may be in pockets initially. In fact, Nokia is already developing mobile phones with IPv6 stack in it!! Initially, a very realistic scenario may be to have small IPv6 networks (within enterprise/metropolitan) and then connect to existing IPv4 backbones. Some mobile operators already have their own infrastructure - this may become IPv6 first.

The opportunity for service companies will be large for migrating existing infrastructure to IPv6 - the question is when is the right time for this? It is definitely not today. I do not foresee a "single D-Day" when the Internet will suddenly and magically become all IPv6 - it will be a very gradual process, following a path similar to the 64 bit operating systems. IPv6 and IPv4 may co-exist for a long time to come!!

Open for Debate: What do you think - Will IPv6 ever become a reality?

References
IPv6 Forum: www.ipv6forum.com
IPng, IPv6 specifications, presentations: playground.sun.com
IPv6 : The Solution for Future Universal networks by Sathya Rao
3GPP Forum: www.3gpp.org
News Article:
3RD GENERATION PARTNERSHIP PROJECT (3GPP) TO ADOPT IPV6.
The 3rd Generation Partnership Project (3GPP) will adopt IPv6 as the protocol for future IP multimedia services. The 3GPP Partnership Project is preparing globally applicable technical specifications for a 3G mobile system based on the evolved GSM core networks and the radio access technologies supported by its partners. - Nokia, May 26, 2000

Sharmila Saha is the Director of MindTree Labs.
She can be contacted at sharmila_saha@mindtree.com