anything, about IPv6: It is an evolutionary step for IP. Calmer heads prevailed during

the two years of IPng working group and IPv6 has become an efficient IPv4 that is

extensible.

IPv4 has proven to be a robust network layer protocol and there have been very few

changes to it over the last 20 years. The biggest problem with IPv4 was the addressing,

and these are the changes that were made. The addressing has not changed, but the

methods of employing the 32-bit addressing have. IPv6 is a direct result of the shortages

of the address space of IPv4. IPv6 is not revolutionary. It is the next step in the

datagram delivery protocol known as IP. It is not a replacement for IPv4 per se, but

there are many new and some revised functions of the protocol that improve upon it.

Currently, there are enough fixes and extensions to the IPv4 protocol (not that there

are many problems with the protocol) to make it last well into the year 2000. I have

heard over the years, why implement a new version of IP when this one is working just

fine? IPv4 simply put a Band-Aid a problem within a time period of need to further

enhance the Internet to reach more people and business requirements.

As you read through this section, you should start to understand that the timing of

this upgrade to IP is about right. The capabilities of IPv6 require a much more

sophisticated computer than was required with IPv4. Generally, IPv4 could run on lowpowered

routers and endstations. The versatility of IPv6 will make use of the higherpowered

routers and workstations.

When we changed IP, we did not change the function of any other protocol—again, the

advantage of modular protocols. TCP and UDP stayed the same. Yes, the software calls

to the IP interface are different: the socket interface known as Berkeley sockets

(Unix), or for PCs the Winsock interface. But the basic functions of TCP/UDP and the

applications that use them are the same. The other protocols that have to change are

those that directly interface with IP. These are Domain Name Server, DHCP, OSPF, RIP,

ICMP, and others.

You will hear a lot about IPv6 over the next few years, and IPv6 implementations will

continue to remain as islands in the IPv4 Internet. This is the correct approach for IPv6.

You cannot “flip the switch” as we did in January 1983 with IPv4. The Internet of today

is extremely large and very commercial. There are still quite a few studies in progress

to determine IPv6 addressing allocation, effects of IPv6 on IPv4 networks, tunneling,

and so on. Slow-but-sure implementation. Test before implementing. Apply applications

that have a need in the marketplace to IPv6. Work out the kinks before

commercialization.

The foundation of IPv6 is IPv4. Like most great things in life, you build upon a

foundation, something that you know works. Cars, over the years, are still built in the

same fashion and still have tires, transmissions, engines, and bodies. But after many

years, the extensions of those basics have led to more than just basic transportation.

Many efficiencies and add-ons have been applied to the basic car to make it safer, better

for the environment, and so forth.

Routers and hosts discovery each other dynamically, hosts can configure themselves

dynamically. There is even a replacement for the DHCP protocol that enforces (and

efficiently uses) IP addressing. And, of course, the biggest change of all for IP: the

address! Placing IPv6-capable nodes on a network with other IPv6 nodes and IPv6

routers will enable an IPv6 network to be established immediately via dynamics.

Neighbor discovery protocols initiate and find the nodes on the network, nodes can

autoconfigure their addresses, and routers simply have to have their interfaces

configured and enabled, and off we go. IPv4 networks prevail, however; probably about

99.99 percent of all networks are IPv4. Therefore, we must make IPv6 work within the

bounds of the existing IPv4 network.

is up and running today, however, through a series of islands that run autonomously

and also use part of the current IPv4 Internet. It is known as the 6Bone and complete

information can be found at:

IPv6 can be grouped into the following categories:

to 128 bits to support more levels of addressing hierarchy, a much greater number

of addressable nodes, and simpler autoconfiguration of addresses. There are three

types of addresses: unicast, anycast, and multicast. The scalability of multicast

routing is improved by adding a “scope” field to multicast addresses. There is no

broadcast address defined.

fields have been dropped and the header is a static 40 bytes.

were experimented with before being presented.

One proposal that had a lot of support wanted to replace IP with the ISO

(Inter-national Organization for Standardization) OSI CLNP Protocol. ISO

CLNP (Connectionless Protocol), which was demonstrated as TUBA (TCP

and UDP over Bigger Addresses. RFCs 1247, 1526, and 1561).

With many changes to the TCP and IP layers, IP version 7 (also known as

TP/IX. RFC 1475) eventually evolved into the CATNIP (RFC 1707).

IP in IP evolved into IPAE (IP Address Encapsulation). It proposed running

two layers of the IP protocol, one for the worldwide backbone and one for

the regional IP networks. This eventually evolved into Simple IP, or SIP.

This moved the address to 64 bits and did away with some of the unused

features of ICMP.

During 1992 and 1993, the Pip internet protocol, developed at Bleacher, was

one of the candidate replacements for IP. It had many improvements in

routing strategies and in mid-1993, Pip was merged with the Simple Internet

Protocol (SIP), creating SIPP (SIP Plus).

SIPP (RFC 1710) is a new version of IP designed to be an evolutionary step

from IPv4. It can be installed as a normal software upgrade in internet

devices and is interoperable with the current IPv4.