IPv6:
The Quiet Revolution
June 9,
2003
By James Klein,
Larta VOX Editor
Internet Protocol version 6 is coming! And has been for a long
time.
Proponents
cite a number of significant improvements provided
by IPv6, but point to a coming shortage of IPv4 addresses
as the primary reason for establishing a new protocol.
Some observers, however, raise issues that could stall
its implementation, including concerns over security
systems, transition costs, control and distribution
of addresses, and a lack of economic incentives, questioning
whether the promise of IPv6 justifies the kind of
resources required to make the switchover at this
time. Whether critics are right or wrong, IPv6 is
expected to gradually replace IPv4 over time, with
the two coexisting for a number of years during a
transition period.
Internet
Protocol version 6 (IPv6) was designed to replace
IPv4, the predominant protocol in use today. A "protocol"
is a naming convention for assigning addresses. Every
computer, server, Internet phone, or any device that
connects to the Internet needs an address. IPv4 provides
a maximum of 4 billion addresses. Supporters of IPv6
insist we need more, and that IPv6 is the only way
to get them. Ipv4 is limited because it uses 32-bit
numbers for its addresses, four octets of 8-bit blocks
(238.14.246.4, for example), while Ipv6's uses 128-bit
sequences, which makes the number of possible addresses
practically limitless.
IPv6
has been called "
one of the most significant
developments in the history of the communications
industry
" by members of the Internet Engineering
Task Force (IETF), which helped develop it, but most
people have never heard of IPv6, don't know it's going
to be implemented, and won't realize it if it is.
Some proponents of IPv6 predict that end users will
run the protocol without even noticing it as companies
incorporate it into their programs.
The
Internet may be a virtual environment, but you still
need an address. Just as every house, apartment, and
business needs an address that is expressed using
a particularly convention, usually involving numbers
and streets and cities and zip codes. Instead of streets
and cities and zip codes, the Internet uses a naming
system, a standard "protocol" that allows
devices to find each other on the Internet. Sometimes
the number of users, however, outstrips the capacity
of the protocol to provide enough addresses. A good
example is the New York City phone system.
In
1930, telephone subscribers in New York City had five-digit
phone numbers.
Colorful two-letter exchange names (BUtterfield, WAverly,
KLondike) were introduced in the 1930s and 1940s to
provide enough telephone addresses for the growing
number of subscribers. Exchange names were widespread
by the 1950s, and many people associate these names
with the period, immortalized in songs and films of
the period, like "Pennsylvania 6-5000" and
"Butterfield-8". Exchanges were phased out
in the 1960s as the area code system was introduced
to accommodate an astronomical increase in the number
of U.S. telephone subscribers. When I grew up in New
York City in the 1970s, it had one area code, 212.
Now it has five.
Not
Your Father's Internet
IPv6
has been in development for about ten years, which
is how long proponents have been issuing dire predictions
about the coming shortage of IP addresses. Even if
IPv4 routing can be scaled to support a full IPv4
Internet, IPv6 supporters warn that the Internet will
eventually run out of network numbers. "There
is no question that an IPng (Internet Protocol Next
Generation) is needed, but only a question of when,"
said Robert M. Hinden of the IETF in a 1995 paper.
When
IPv4 was originally implemented, no one imagined that
IP addresses would become scarce, but in the 1990s,
many predicted the steadily increasing demand for
addresses would exceed the ability of IPv4 to supply
them. The IETF first raised concerns about a possible
shortage of IPv4 addresses in 1992, and began an IPng
program to determine the best solution. IPv6 was recommended
by the IPng group and approved by the IETF in 1994.
Since then, groups including the IETF, the Internet
Society (ISOC), and IPv6 Forum have been relentlessly
beating the drum for a switch to IPv6. Some companies
and individuals have heeded the call, others have
put their fingers in their ears, and others have said
they simply don't like the beat.
In March of 2002, however, the IETF put a temporary
halt on the development of all IPv6 migration tools
so it could get a handle on the technology before
a slew of standards caused problems. While the technology
is stable today, IPv6 is not ready for worldwide adoption
because of all the migration tools required, at least
according to Margaret Wasserman, IETF next-generation
and IPv6 transition working groups co-chair. As reported
by Jim Wagener in internetnews.com, the IETF still
needs to test all the tools before they will give
their seal of approval, and called the development
halt so they could at least address the tools already
made. Critics say the development shutdown stifles
innovation. The IETF has indicated that it will open
up IPv6 for tool development this summer and expects
worldwide IPv6 adoption sometime in the next couple
years.
Has the well run dry?
Whether
or how soon you believe the Internet will run out
of addresses depends who you ask. Proponents of IPv6
say the Internet is a victim of its own success, and
predict a global shortage of addresses within a couple
of years. These observers fear devices soon won't
be able to use the Internet at all, and see a time
when the lack of IP addresses will stifle technology
development and economic growth. Others expect it
may take a decade or so before the number of devices
connected to the Internet outstrips the supply of
addresses, and say it's better to wait and see what
happens in the near future before driving a switch
to the new protocol.
The
number of Internet users is expected by many to grow
exponentially, particularly in Asia and developing
countries. In addition, the number of IP addresses
each user requires is projected to increase as new
technologies emerge. New generation mobile phones,
high-speed data access, wireless voice-over-Internet
devices, peer-to-peer applications, as well as Internet-equipped
vehicles and household appliances will all need individual
IP addresses, multiplying the number of addresses
required by each user. According to the U.S. Census
Bureau, the world may be inhabited by as many as 9
billion people by 2050. If you count all personal
and business communication and information systems,
including vehicles and home appliances, each of those
9 billion people might be using dozens of devices,
each with its own address, that connect to the Internet.
Besides
providing more IP addresses, IPv6 offers many other
benefits. IPv6 makes routing and automatic configuring
easier, simplifies header formatting, improves support
for extensions and options, increases flow labeling
capability, provides for greater security, and consolidates
authentication and privacy capabilities. IPv6's simplified
header architecture and protocol operation could reduce
operational expenses, but perhaps the most significant
improvement offered by IPv6 is its address autconfiguration
capabilities, which allows mobile devices to be quickly
recognized as they move among networks. IPv6 is also
friendlier to traffic engineering technologies, has
mandatory multicast specifications, better support
for ad-hoc networking, and other technical compared
to IPv4. Even so, the greatest benefit of IPv6 continues
to be its ability to provide a practically inexhaustible
number of addresses.
Asia
Left Out
Asian
countries have the most to lose if IPv4 runs out of
addresses, and therefore include some of the governments
that are most active and vocal in supporting the switch
to IPv6. China and India have already run out of IP
addresses and are migrating to IPv6, with or without
worldwide approval. Some predict there may come a
time when there will be a "dual-track Internet"
on either side of the Pacific Ocean, as Asia moves
rapidly towards IPv6 while North America lags behind.
When
IPv4 addresses were originally allocated, about 20
years ago, the U.S. and Europe grabbed most of them.
The U.S. controls over 3 billion IP addresses, which
represents about 70 percent of all IP addresses available.
Asian countries were largely excluded from the address
land rush. China and South Korea, for example, were
allocated only about 38 million and 23 million IP
addresses, respectively. China has fewer IPv4 addresses
available than Sony Group. Last year, China had about
17 million Internet subscribers, while some project
the figure to reach 62.5 million by 2007. In Japan
today, current IPv4 address allocation is insufficient.
China
and Japan are investing millions to develop IPv6.
Both governments have pledged $32 million dollars
to invest in network construction and testing, system
development, application technology development and
standardization. The Nikkei Electronics news service
has reported that the Japanese firm Hitachi will become
an Internet service provider (ISP) in China, and will
be the first in the country, and probably the world,
to use only IPv6 addresses for customers, although
non-Asian makers such as Nortel, Cisco, Nokia and
others have supported IPv6 in their products for some
years. Japan has led the world in IPv6 implementation,
especially in its use of protocol stacks.
"We anticipate that this demand [for IPv6] will
be strongest in the Asia-Pacific region, which owns
far fewer registered IP addresses than does North
America. As Internet-accessible wireless devices and
Internet use in general grow in Asia-Pacific, service
providers targeting that region will seek to build
IPv6 backbone networks. Gartner expects that by 2006,
50% of all carriers in this region will have IPv6
running in some portion of their networks," stated
Lawrence Orans, senior analyst, Gartner, Inc., in
a 2001 Cisco News article.
Supporters of IPv6 also raise humanitarian concerns,
describing a worldwide digital divide between those
countries, businesses and individuals who are able
to acquire Internet addresses and those who can't,
resulting in a shortage of IPv4 addresses that could
hinder economic development in poorer nations.
Because
so many issues have to be worked out in order to facilitate
the switch to IPv6, temporary fixes were developed,
such as Network Address Translation (NAT), to prevent
a shortage of IP addresses. NAT allows a number of
hosts to share a single IP address. Proponents of
IPv6 emphasize the limitations of NAT technology,
especially when used in peer-to-peer networks. NAT
also presents challenges for applications that incorporate
the host's IP address in the application-layer data,
an issue particularly problematic for security protocols
such as IPSec.
A
Foregone Conclusion?
"We
view IPv6 as inevitable," said Andy Ingram of
Sun Microsystems, which implemented a dual stack IPv4-IPv6
architecture in the first release of its Solaris 8
system. A paper written by members of the Internet
Engineering Task Force (IETF) stated that "The
first (and most important) fact is that IPv6 is inevitable."
Whether the critics are right or wrong, IPv6 may have
gathered enough momentum to make it unstoppable. The
ranks of IPv6 supporters have swelled to operatic
proportions with a chorus of voices that includes
groups like IETF, ISOC, and the IPv6 Forum, as well
as ISPs, switching and router companies, and hardware
and software producers.
HP,
Cisco, NEC, Nokia, NTT, Ericsson, Sony, Hitachi, British
Telecom, Matsushita, Sun Microsystems, Microsoft,
Juniper, and many other companies have spent millions
of dollars testing IPv6 technologies and integrating
IPv6 capabilities into their products and services.
IPv6 implementation programs have been developed for
many different host operating systems and routers,
including implementations by Apple, BSDI, Bull, Digital,
Epilogue, FreeBSD, FTP Software, Hitachi, HP, IBM,
INRIA, Interpeak, Linux, Mentat, Microsoft, NetBSD,
Nokia, Novell, NRL, NTHU, OpenBSD, Pacific Softworks,
Process Software, SICS, SCO, Siemens Nixdorf, Silicon
Graphics, Sun, UNH, and WIDE, and router implementations
by 3Com, 6WIND, Bay Networks, Cisco Systems, Digital,
Hitachi, IBM, Merit (routing protocols), Nokia, NTHU,
Sumitomo Electric, and Telebit Communications. Microsoft
has been contributing to the IPv6 standardization
effort since 1996. The Internet Assigned Numbers Authority
(IANA), which is responsible for allocating Internet
addresses has started issuing numbers based on IPv6.
Sony plans to start shipping IPv6-enabled products
in the fall of 2003.
The
Internet needs an upgrade
Upgrades.
You love them, you hate them, you can't do without
them. Or can you? Sometimes I'm delighted to find
all the wondrous new features in an upgrade and do
bless the wise and benevolent seers who improved what
I now realize was a shoddy and ill-equipped piece
of software. More often, however, I am enraged as
I'm forced to relearn applications I already mastered.
The new versions are usually more complicated than
the previous versions, which makes it even more fun.
Simple commands are seemingly changed at random. I'm
still irked whenever I use an activated table of contents
in Microsoft Word as they changed the command from
a simple, one-handed click of the mouse to an awkward
combination of the control key with one hand and the
mouse with the other. It got so that I wouldn't return
email from my IT team when they issued requests to
upgrade my computer. I'd eat at my desk for fear of
running into them in the cafeteria. But proponents
of IPv6 insist the transition will be smooth and untroubled.
Supporters claim IPv6 can be installed as a normal
software upgrade in Internet devices and is interoperable
with the current IPv4. Its deployment strategy is
designed to not have any flag days or other dependencies.
IPv6 is designed to run well on high performance networks
and will also still be efficient for low-bandwidth
wireless networks. In addition, IPv6 provides a platform
for new Internet functionality that will be required
in the near future.
IPv6 includes a transition mechanism that is designed
to allow users to adopt and deploy IPv6. In addition,
the Simple Internet Transition (SIT) is a set of protocol
mechanisms implemented in hosts and routers, along
with some operational guidelines for addressing and
deployment, designed to make transitioning the Internet
to IPv6 work with as little disruption as possible.
SIT provides for an incremental upgrade and deployment
of IPv6, with minimal upgrade dependencies, easy addressing,
and low-cost start-up costs.
Not
Everyone's Convinced
The
primary reason given for supporting IPv6 is an expectation
that the current address naming protocol, IPv4, will
soon run out of available addresses, but a stagnant
global economy and an excess capacity of bandwidth
and networking equipment compel some critics to question
whether IPv6 is as crucial as it was once thought
to be.
Whether
or not the new protocol is required at all has been
the subject of some debate within the technical community.
While some argue that IPv6 is required for the future
growth of the Internet, there remains a sizable camp
who argue that the protocol is "too little, too
late" and that IPv4 with the addition of network
address translation (NAT) offers a viable system for
the future.
What
proponents of IPv6 sometimes fail to mention is that
the number of IP addresses actually used is very low
compared with the theoretical total. IP addresses
are used by a small number of the Earth's inhabitants.
Temporary solutions have been able to fix most of
IPv4's problems, and address its limitations, while
some critics say IPv6 doesn't fix the system's biggest
routing problems. Others have called IPv6 "a
full employment act for software engineers."
Vint
Cerf, the honorary chairman of the IPv6 Forum, has
stated that most businesses are "pretty near"
apathetic toward the problem of address depletion,
but has also said that proponents faced similar challenges
when they encouraged the transition from Arpanet to
today's Internet, which also required administrators
to overhaul their systems.
George
Jetson in China
One
of the rationales for IPv6 is the expectation that
we'll all have numerous devices connected to the Internet
- computers, printers, digital cameras, radios, IP
phones, Internet-enabled toasters and web-connected
robotic house pets. One of IPv6's most lauded features
is its autoconfiguration capability, which is especially
useful for consumers with multiple devices. While
high-end consumers, computer geeks, Sharper Image
shoppers, and George Jetson wannabes may surround
themselves with web-connected paraphernalia, it is
unlikely that billions of Chinese consumers will be
able to afford all these gadgets any time soon, which
undermines the figures given as the primary justification
for IPv6.
Band-Aids
IPv6
has been delayed in part because temporary fixes such
as network address translators and dynamic addressing
have resolved the potential lack of IP address, while
dual-stack approaches, which allow IPv4 and IPv6 to
run on the same device, lessen the need for immediate
implementation. While most agree temporary solutions
are not as robust, feature-rich or comprehensive as
IPv6, critics believe they may provide adequate solutions
until a better plan comes along, but IPv6 supporters
have warned that continuing to add temporary solutions
to IPv4 will only encumber a system that wasn't designed
to be a worldwide network in the first place.
NAT
has been so successful that some believe IPv6 is unnecessary,
but while NAT can be used to cleverly hide high numbers
of private addresses behind a much lower number of
public addresses, NAT has a number of performance
and management problems that could cause problems
for some applications, including many peer-to-peer
networking programs. IPv6 proponents point out that
NAT inhibits many forms of innovative network use,
and presents very challenging operational problems
when deployed on a global scale.
Some
observers speculate it may be possible to reclaim
addresses from organizations that no longer need them.
However, this solution may require manual renumbering
of address networks, which could be costly. A few
organizations, such as Stanford University, have returned
addresses, but this practice has yet to gain much
momentum.
Economic
Incentives Lacking
While
most observers admit IPv6 has advantages that will
improve business on the Internet, skeptics have raised
economic issues in their arguments against IPv6, and
even proponents admit that a lack of IP addresses
alone is not necessarily a strong enough reason to
drive the transition to IPv6. IPv6, an Internet Evolution,
a paper by the IPv6 Forum, which has "a clear
mission to promote IPv6", states that "While
the primary reason for the next generation protocol
was increased address space, this alone is insufficient
to drive an economically sounds transition strategy."
The
IPv6 Forum paper explains that IPv6 may be able to
be used to create new revenue streams, but doesn't
detail what those revenue streams will be, and recognizes
that IPv6 is essentially a software upgrade, which
will require users to learn and relearn certain features.
Cost reductions are cited in the paper, but transition
costs are equally significant: "
there is
a substantial increase in the theoretical spending
for IT (with the adoption of IPv6), which can be directly
attributed to the transition costs (new equipment,
application upgrades, training)
" The paper
goes on to admit that adoption of IPv6 among Internet
service providers has been lackluster. "For the
ISP, particularly in North America, adoption of IPv6
has generally not been a hot topic. The potentially
high cost of transition for the ISP, accompanied by
a serious lack of business drivers, has not precipitated
a great deal of interest of the telecoms."
The
issue of transition costs was raised as early as 1999.
A CNET news article from 1999 addressed concerns that
the costs of implanting IPv6 may outweigh the benefits:
"
those pushing adoption of the protocol
face tremendous obstacles because of the expenses
businesses would incur to make the change. With IPv6,
businesses would have to make massive infrastructure
changes to their networks, which are costly and complicated."
"What
I'm desperately trying to get the [Internet Engineering
Task Force] to understand and face up to is that the
solution they put forward is not going to be acceptable
in the marketplace," said Noel Chiappa, a networking
consultant and critic of IPv6, in the 1999 CNET news
article.
Other
critics have pointed out the lack of new killer apps
that could whet consumers' appetites for IPv6, while
others speculate that the new protocol system will
open the door for many applications that haven't yet
been created. Whether new applications emerge or not,
IPv6 will likely provide greater efficiency running
existing and future programs on the Internet.
Some
observers have likened the dilemma faced by proponents
of IPv6 to that faced by some engineers several years
ago when they tried to warn governments and businesses
of the Y2K problem. Even though the potential economic
costs were astronomical, many businesses did nothing
in the face of a potential looming disaster because
the solutions were prohibitively complicated and expensive.
Say
You Want a Revolution
Other
issues have been raised in prelude to widespread IPv6
adoption. One issue involves the way address registries
distribute IP address blocks to service providers
and major customers. As history has proven, revolutions
sometimes result in unexpected power-grabs as new
leaders are absolutely corrupted by absolute power.
Some critics of IPv6 have raised concerns about exactly
who will be controlling the new addresses, how they
will be distributed, and how privacy will be protected,
given that the current system in place hasn't treated
all countries fairly. Others warn that the same wasteful
tendencies that plagued Ipv4 will affect Ipv6 if the
same unfortunate policies are pursued, contending
that while IPv6 has a seemingly limitless amount of
address space, short-sighted and wasteful allocation
policies could result in the adoption of practices
that lead to premature exhaustion of the IPv4 address
space.
IPv6
address numbers will be assigned and reassigned according
to practices implemented by Internet Registries, who
distribute IP address space to their members and customers.
Local Internet Registries are overseen by National
and Regional Internet Registries, who are ultimately
controlled by the IANA.
Some
observers have also raised privacy issues that should
be addressed before full implemented of IPv6 takes
place. Part of the IPv6 address is sometimes derived
from the LAN address of the computer's network interface
card, which could be misused to identify for the computer
behind the IP address. It is feared that market researchers
could use data-mining techniques to match addresses
to individuals. The IETF has addressed this issue
by developing an alternative solution in which the
fixed identifier is replaced by a randomly generated
number.
Some
critics also describe flaws in routing directories
that help servers on one part of the Internet find
addresses on other parts. These routing tables were
not designed to handle large volumes of traffic, and
could experience problems given the steady growth
of Internet usage. IPv6 does nothing to alleviate
this problem.
As
the debate continues, it's clear that IPv6 proponents
would do well to convince more people of the advantages
of IPv6. Whether they do or not, however, IPv6 may
well become the standard for Internet addressing in
the future.
Return
to this week's issue of LA VOX >