www.SatelliteTCP.com Page 58 Copyright © DC Palter
Chapter 8: IPv6
The current version of the Internet Protocol, IPv4, was designed assuming it would be used by a
relatively small network of engineers and scientists, primarily transferring files and exchanging
e-mail. IPv4 was developed more as a computer science research project than as the backbone of
a critical global infrastructure underpinning communications for consumers, businesses,
governments, and militaries.
In the early 1990s, as growth of the Internet exploded and TCP/IP became the de facto standard
for computer networking, the IETF community realized that IP had some deficiencies that
needed to be corrected to make it better suited to its growing role. The IETF therefore began
development of a new version of IP, designated IPv6.
The most pressing issue at the time was the anticipation that the IPv4 address space would be
fully exhausted by 2005 or earlier, especially if plans to network all sorts of devices from home
appliances to wearable sensors come to fruition. IPv6 therefore increases the size of addresses
from the 32 bits used by IPv4 to 128 bits, which is said to be sufficient to separately address
every grain of sand on earth. (Presumably, at the time that interplanetary exploration makes it
necessary to network every grain of sand on additional planets, we will develop IPv7.)
However, since it was impossible to squeeze a 128-bit IPv6 address into the existing IP header,
backward compatibility between IPv4 and IPv6 was impossible. Because a completely new
header format would be required anyway, the IETF decided to completely rewrite IP to add
additional enhancements at the same time to make the protocol more scalable and efficient.
8.1 Benefits of IPv6
The most obvious benefits of IPv6 revolve around the expanded address space. The larger
addresses not only allow continued expansion of the number of users on the Internet, it can also
be used to improve router efficiency by aggregating addresses into route hierarchies. In addition,
the larger address space makes it possible to translate Ethernet, IPX, and other non-IP addresses
into IPv6 addresses.
IPv6 includes a variety of features to improve the efficiency of routers processing IPv6 packets.
For example, the packet header is simplified from 12 data elements in IPv4 to only 8 elements in
IPv6. IP options are moved into separate extension headers which usually only need to be
examined by the end nodes, reducing the amount of processing required by the intermediate
In addition, many features which are optional in IPv4 are required in IPv6, guaranteeing the
ability to take advantage of:
www.SatelliteTCP.com Page 59 Copyright © DC Palter
8.2 Transition Mechanisms
Of course, with IPv6 not being backward compatible with the existing IPv4 infrastructure,
mechanisms are necessary to operate during the transition period from IPv4 to IPv6. Unlike
previous fundamental changes to the networking protocols where a “flag day” required all users
to switch protocols, the current installed base of IPv4-based PCs, applications, and network
infrastructure makes this type of abrupt transition entirely impossible now.
Instead, transition mechanisms focus on being able to use both protocols during the long period
until we will be able to assume that all devices and applications will be running only IPv6 over
an IPv6 network. These transition mechanisms include:
IPv4 address translation: IPv4 addresses can be translated into IPv6 address by adding a
prefix of leading zeroes.
Dual protocols: For the foreseeable future, all implementations of IPv6 will also include the
ability to run IPv4.
IPv6 tunneled over IPv4: IPv6 end nodes can communicate with each other through IPv4
routers by encapsulating IPv6 packets within IPv4 packets.
8.3 IPv6 Experience
Although the designers of IPv6 envisioned a fairly lengthy transition period, the transition has
been much slower than anticipated. At the time of this writing, several years after the IETF
began the standardization process and vendors were demonstrating interoperable
implementations of IPv6, actual deployment is still limited to a small number of research
There are a few reasons why deployment of IPv6 has been slow. First, the pace of IPv4 address
exhaustion has slowed considerably due to the use of NAT to hide whole networks behind a
single IPv4 address, while reclamation of some of the Class A and Class B addresses that were
handed out in the early days of the Internet and were mostly unused has freed up additional
addresses. Meanwhile the switch from classful to classless addressing has reduced the explosion
of separate routes in the backbone while the increasing capability of high-end routers has mostly
nullified the issue of router efficiency anyway.
Having sufficient IPv4 addresses available to end-users has created a chicken-and-egg situation.
With no need for end users to transition to IPv6, ISPs and network administrators have no
incentive to spend the considerable time and effort to re-engineer their entire network to support
IPv6. With no networks using IPv6, there is little point for end users to switch to IPv6 and no
need for software vendors to revise their applications to support IPv6 addressing.
www.SatelliteTCP.com Page 60 Copyright © DC Palter
IPv6 developers originally hoped that some of the advanced features of IPv6 would outweigh the
pain of the transition effort, but instead, IP Security, Path MTU, and DHCP have became widely
available as part of IPv4.
Still, the IPv4 address space is limited and will run out eventually. The 3G mobile phone
network, anticipating millions of users with an IP address for each phone, will require the use of
IPv6. It is also likely that countries such as China which were late to the Internet boom and
ended up with an address allocation that can only support a small fraction of the population will
need to transition to IPv6 sooner than countries such as the United States which may not need
additional addresses for many years.
8.4 IPv6 Datagram Format
Figure 22: IPv6 Packet Format
Version IP version number, set to 6.
Traffic Class Traffic class analogous to IPv4 Type of Service field.
Flow Label A single connection or other flow can be marked so that each packet receives
the same routing or special treatment, allowing the router to avoid having to
analyze the rest of the header. Similar to MPLS.
Payload Length Specifies the length of the datagram following the header.
Next Header Identifies optional header fields. The optional fields are usually for the end
nodes and do not need to be processed by the routers.
Hop Limit Same as the IPv4 Time-to-live field.
Source Address 128-bit address of the originator of the packet.
www.SatelliteTCP.com Page 61 Copyright © DC Palter
Destination Address 128-bit address of the intended recipient of the packet.
8.5 IPv6 and Satellites
It is a common misconception within the satellite industry that IPv6 will overcome the
limitations of TCP/IP over satellite. This is entirely wrong. IPv6 revises the IP layer. However,
IPv4 works fine over satellite and none of the changes of IPv6 will have any effect on how IP
runs over satellite. The limitations of TCP/IP over satellite are strictly due to the design of TCP,
and IPv6 does not change the operation of TCP.
IPv6 is the next-generation version of IP. The primary difference between IPv4, the current
generation of IP, and IPv6 is the increase in the source and destination addresses from 32 bits to
8.7 Further Reference:
RFC 1705: Six Virtual Inches to the Left: The Problem with IPng (informational)
RFC 2460: Internet Protocol, Version 6 (IPv6) Specification (draft standard)