A Tale of Two Protocols

IPv6 Glossary

6Bone: The 6Bone is a test network for IPv6, based on a worldwide backbone connecting sites in over 42 countries which run IPv6 implementations from some 30+ vendors and institutes.

Broadcast Addressing: Addressing all interfaces on a network; usually provided as a feature of the physical structure of a LAN, such as Ethernet

DNS: Domain Name Service. In a nutshell, these are the name equivalents for IPv4 addresses.

Field: A contiguous grouping of information, as in a packet header

Flag Day: An event when operation of a system or systems must be suspended to convert to a new mode of operation or software (an undesirable situation)

Fragmentation: Breaking a large packet into smaller units for transit over a size-restricted transmission link (e.g., when a packet's size exceeds the Maximum Transfer Unit [MTU])

Header: The first part of a packet describing its content and processing options

NAT: Network Address Translation. Computers and/or software that translates older protocols for newer networks.

IDRP: Inter-Domain Routing Protocol - for IPv6 routing between subdomains.

IETF: Internet Engineering Task Force.

Internet Protocol: A data communications protocol for transfer of information in packets by means of a best-efforts, connectionless, unreliable (non-error-correcting scheme)

IPng: IP next generation - a term used to refer to IPv6 until the version 6 designation was official.

IPv4: Internet protocol developed in the mid-to-late 1970s that gives us the foundation for the current Internet. (Eg.,

IPv6: Next-generation Internet protocol. Current version is IPv4.

MTU: Maximum transfer unit; limit on packet size able to transit a physical link

Multicast Addressing: A feature of a network that can automatically replicate the server's packets and route them to each subscriber in a group. A similar parallel is found in Cable Television.

Protocol: A well-defined set of formats and commands to manage the transfer of information.

RIP: Routing Information Protocol; a routing protocol within subdomains, easily implemented but not able to make efficient use of connections.

SIT: Simple Internet Transition. Rules and protocols to simplify migration from IPv4 to IPv6.

TCP: Transmission Control Protocol; a higher-level protocol using the services of IP to provide reliable connection-oriented services for applicatoins (e.g., Telnet, World-Wide Web, electronic mail, etc.)

TLA: Top-Level Aggregation Identifiers. Bit boundaries are divided by TLAs. A "sub-TLA" area of 13 bits, with a reserved 6 bits, has been designated for initial allocations, which will be made on a "slow start" of a sub-TLA, e.g., a /35 block.

Unicast: Addressing a single interface for packet transmission.
Source: From the Net

Written by: Avinash

he long-held promise of anywhere anytime access is no longer confined to the realms of a tantalising fringe possibility, as the next generation of mobile technology gets set to replace the PC. The demand for network intelligent devices such as mobile phones, handhelds and home area networks is fuelling the need for Internet addresses exerting intense pressure on existing IP protocols such as IPV4. The IPV4 address space system developed as an extension for a US defence project 20 years ago is falling woefully short of address space and is not robust enough to support services such as realtime traffic, flexible congestion control schemes and easy to use security features.

Says Mr. Rinka Singh, Senior Manager at Novell, "A major problem faced by the existing IPV4 protocol is the unending demand for Internet addresses created by the massive upsurge in computing that we have witnessed in the course of the last decade and a half. The rapid growth of mobile internet applications and new generations of mobile phones have further stimulated a demand for huge numbers of extra IP addresses. As a matter of fact 70 per cent of the addresses have already been consumed and as far as the remaining 30 per cent is concerned a lion's share of it cannot be used due to various factors. So we have reached a stage where we actually are running out of addresses".

The existing manifestation of TCP/IP uses a 32-bit address space, which theoretically makes it possible to have only 4,294,967,296 addresses. Experts such as Singh claim that the actual number is far smaller because a huge share of the TCP/IP address space is used for internal addressing and cannot be used as address space. With the Worldwide Internet users touching 553 million people in August 2002, according to a Global Internet Trends report from Nielsen//NetRatings, it would create an equivalent demand for addresses. As the shortcomings of the older protocol become glaringly evident, the enduring solution to the address space crunch was to devise a new addressing scheme. The IETF has designed IPv6, which tides over the address shortage by using a 128-bit addressing scheme, making it possible to assign over 3.4 x 1038 individual addresses in the whole network. IPv6 thus creates a virtually inexhaustible number of addresses for every person and machine on the planet.

The new IPV6 protocol is also superior in terms of other functions. "A persistent problem, that has started to afflict TCP/IP in the last several years is the deficiency of end-to-end security in the protocol. There are optional add-ons that provide security in certain contexts, such as IPSec or HTTPS, but they are not built into the TCP/IP itself, and thus not implemented uniformly." says Hemanth T D, President of IPV6 India Forum.

As a large number of enterprises have experienced firsthand to their utter discomfort, the last few years have seen a dramatic rise in distributed, difficult-to-trace attacks which have exposed TCP/IP's inherent lack of security. Thus, people today are finding it harder to rely on IPV4 for end-to-end security. Hemanth adds that in sharp contrast to IPV4, IPv6 supports IPSec intrinsically as a result of which security features such as packet encryption and source authentication are guaranteed. The bottom line is, in the case of IPV4, IPSec was optional; in IPv6, it is mandatory.

Another bane of IPV4 is that it requires manual configuration of addresses which is cumbersome. In contrast IPV6 provides a vigorous auto configuration formula, through which an interface can configure itself with an address by using router discovery to determine the best possible address.

The Internet's routing tables by and large contains 80,000 entries per router node. IPV6 provides an efficient routing infrastructure to handle the sheer amount of connections involved. According to Hemanth, "Quality of service, which is a way of insuring that certain streams get priority handling doesn't exist in IPv4 whereas this feature is endemic to the system in IPv6. Hemanth further drives the point home by pointing out that IPV6 features like discovery of other local nodes and interfaces, and extensibility for adding new headers have gone a long way in making the Internet more efficient."

Despite the robustness of the new protocol there are several roadblocks on the way of IPV6 adoption. The primary being there is no single road to IPV6 migration or transition to the interoperation of IPV6 and IPV4. People have devised a few stopgap arrangements to get over this address deficiency predicament. The curtailmet of technology spending by companies is delaying the process of migration. The attitude if it ain't broke why change it not speeding deployment of the new protocol. According to Hemanth, "A major problem, which is endemic to countries like India, is the sheer lack of awareness of the functionalities that IPV6 brings." Currently the problem of delayed migration are being met by using Network Address Translators (NAT) to assign temporary addresses to transient nodes. Another idea that is in practice is using host headers to support multiple DNS entries on a single IP address. But the downside to these stopgap arrangements is that they bring with them their own baggage of technical issues like the lack of encryption and other advanced network Also IPv4 cannot continue to be patched up with the likes of NAT, as it doesn't solve the problem of a lack of global IP addresses.

Another major roadblock is that the transition to IPv6 requires that every operating system that connects to the Internet should be able to support it. Ideally, operating systems should be able to support both IPv4 and IPv6 at the same time, since the two networks will coexist for some time before the transition is total. "It is a paradox that the networking industry's ability to produce NAT, CIDR, and other stopgap arrangements to overcome IPv4's shortcomings has become one of the major roadblocks to IPv6" says Mr. Rinka Singh. According to him, enterprises will realize that, in the long run NAT's address translation architecture makes it difficult to implement end-to-end packet-level security in transactions.

Despite these drawbacks, efforts are on to have IPv6 rolled out on the Internet either through creating closed private networks to test IPv6, or semi-public networks, which are open to people who have both the right equipment and software. According to Hemanth, several leading software companies in Bangalore are taking a shot at IPV6 through Bangalore V6 (BV6) on which several IPV6 compliant applications are run.

"IPV6 will be up and running in India within a matter of 18-24 months provided the government gives the much needed push. The Japanese and Korean governments have mandated its deployment within their countries," asserts Mr. Rinka Singh. Mr. Singh adds that the Japanese and Korean governments have mandated its deployment within their countries. The European Commission has also advised its members to migrate to networks supporting to IPv6 by 2005. Thus, in spite of the initial teething problems IPV6 is gradually becoming the protocol of choice the world over. In Europe, the first commercial IPv6 service network, Telia/Skanova, was put in place last year using gigabit routers. These routers can handle IPv4 and IPv6 traffic at full speed. Other carriers have gone through evaluations and initial trials with IPv6 products and technology, and it's estimated that by early to mid-2003 they will be complete and will then be ready for full deployment soon after.

The emergence of IPV6 has thrown open a new set of opportunities for a wide gamut of applications. The security standards that IPV6 will bring to the table will give a huge boost for E-business. Another area where IPV6 promises a great deal is mobile wireless communication. A basic precept of next-generation wireless service is constant access to the Internet. Such 24/7 connections will obviously require a fixed IP address. Thus next-generation wireless service will kindle a huge demand for IP addresses, which thankfully, IPV6 will be in a position to provide. IPV6 will also support wireless location devices that could be used to track shipping containers, trucks etc. In the near future IPV6 will enable commonplace devices such as streetlights, fire alarms, refrigerators etc to turn intelligent by sporting IP addresses. Thus, the applications for IPV6 are frankly limitless.

(Avinash is a freelance writer, who has contributed to various publications.)

Email this article | Respond to this article