Cost Benefit Analysis of implementing IPV6

INTRODUCTION
This report focuses on one of the communications protocols that define the infrastructure of the Internet
the so-called Internet Protocol (IP), which enables data and other traffic to traverse the Internet and to
arrive at the desired destination. IP not only provides a standardized envelope! for the information that is
sent" it also contains headers! that provide addressin#, routin#, and messa#e-handlin# information that
enables a messa#e to be directed to its final destination over the various media that compose the Internet.
The current #eneration of IP version $ (IPv$), has been in use for more than %& years and has supported
the Internet's #rowth over the last decade. (ith the transformation of the Internet in the )**&s from a
research networ+ to a commercialized networ+, concerns were raised about the ability of IPv$ to
accommodate emer#in# demand, especially the anticipated demand for Internet addresses. ,s a result, an
international or#anization, the Internet -n#ineerin# Tas+ .orce (I-T.), be#an wor+ on the ne/t #eneration
IP. Its efforts led to the development of IPv0.
IPv0 will enable an enormous increase in the number of Internet addresses currently available under IPv$.
1emand for such addresses will increase as more and more of the world's population re2uest Internet
access. 3isco 4ystems notes that if the )5 lar#est countries were to assi#n uni2ue addresses to only %&
percent of their populations, the resultin# demand would easily e/haust the remainin# supply of IPv$
addresses. 3ontinued #rowth in mobile data services via wireless telephones and data terminals, such as
personal data assistants (P1,s), will also e/pand demand for Internet addresses. The situation may become
critical if, as some pro6ect, a mar+et emer#es for in-home devices (e.#., smart appliances! and
entertainment systems) that are accessible from outside the home via the Internet. ,lthou#h considerable
disa#reement e/ists as to whether, to what e/tent, and at what pace, such demand will develop, it is
e/pected that deployment of IPv0 would provide the address space to accommodate whatever level of
demand does emer#e.
7esides affordin# e/ponentially e/panded address space, IPv0 has been desi#ned to provide other features
and capabilities, includin# improved support for header options and e/tensions, simplified assi#nment of
addresses and confi#uration options for communications devices, and additional security features.
1evelopment of IPv0, moreover, has stimulated enhancements to IPv$. ,s useful capabilities have been
devised for IPv0, protocol developers and manufacturers have wor+ed to incorporate a number of those
same capabilities into IPv$. ,s a result, IPv$ can now support, to varyin# de#rees, many of the capabilities
available in IPv0. ,t the same time, additional mechanisms and tools have been developed to miti#ate, to an
e/tent, the IPv$ address e/haustion concerns that in lar#e part prompted the development of IPv0.
There is a massive embedded base of IPv$ e2uipment and applications in the communications system we
+now as the Internet. The capabilities of IPv$, which have been enhanced over time in response to the
development of IPv0, ma+e IPv$ functionality sufficient to serve the needs of many current Internet users
and service providers. 3onse2uently, an important policy 2uestion concernin# IPv0 deployment is whether
the incremental benefits of adoptin# IPv0 6ustify the costs of convertin# the lar#e embedded IPv$ base to
IPv0 on an accelerated basis (e.#., well in advance of an or#anization's normal e2uipment replacement
cycle).
7ecause of conversion costs and the comple/ities involved in predictin# a return on investment for IPv0 in
the short term, most observers believe that there will be a considerable transition period durin# which IPv$
and IPv0-based networ+s will coe/ist. 1urin# that transition, firms will incur costs to ensure interoperability
amon# e2uipment, applications, and networ+s, both domestically and to a lesser e/tent internationally.
4imultaneous operation of IPv$ and IPv0 may also re2uire additional effort to ensure communications
security and to protect networ+s from attac+. These transition costs, in addition to the more obvious direct
costs of convertin# to IPv0 and ma+in# any other necessary networ+ chan#es, should be considered when
assessin# the benefits of adoptin# IPv0. -nterprises must determine whether the cumulative benefits of
deployin# IPv0 will 6ustify the costs of mi#ratin# from IPv$ to IPv0.
BENEFITS AND COSTS OF ADOPTING IPv6
Industry sta+eholders and Internet e/perts #enerally a#ree that IPv0-based networ+s would be technically
superior to IPv$-based networ+s. The increased address space available under IPv0 could stimulate
development and deployment of new communications devices and new applications, and could enable
networ+ restructurin# to occur more easily. The redesi#ned header structure in IPv0 and the enhanced
capabilities of the new protocol could provide si#nificant benefits to Internet users, networ+ administrators,
and applications developers. IPv0 could also simplify the activation, confi#uration, and operation of certain
networ+s and services.
8any of the prospective benefits of IPv0, moreover, appear to be predicated on the removal or modification
of middlebo/es! that affect direct Internet communications between end-user devices, such as 9etwor+
,ddress Translation (9,T) devices (see 4ection %.).).%), firewalls, and intrusion detection systems (I14). It
remains to be seen whether or when such devices will be either phased out or made transparent to end-to-
end (-%-) Internet communications and applications.
In this section, we discuss the benefits and costs of adoptin# IPv0. ,fter first evaluatin# the potential
benefits of deployin# IPv0, we discuss the nature and relative ma#nitude of the costs that enterprises and
individuals may incur to deploy IPv0.
Relative Benefits of IPv6 vs. IPv4
, #eneral consensus appears to e/ist re#ardin# the technical improvements of IPv0 versus IPv$ and the
types of benefits that could follow from widespread adoption of IPv0. 1isa#reement e/ists, however,
re#ardin# the size of those benefits and whether the incremental benefits of IPv0 (versus IPv$) for some or
all users would outwei#h the costs of a #reatly accelerated transition from IPv$ to IPv0
, principal by-product of deployin# IPv0 would be a lar#e increase in the number of available IP addresses.
The :%-bit address field in the IPv$ pac+et header provides about $ billion ($/)&*) uni2ue Internet
addresses. The )%;-bit address header in IPv0, in contrast enou#h to assi#n addresses to each person now
on earth or even to every s2uare inch of the earth's surface.
The vast pool of addresses available under IPv0 would, at a minimum, <future proof< the Internet a#ainst
potential address shorta#es resultin# from the emer#ence of new and unforeseen services or applications
that re2uire lar#e 2uantities of #lobally routable Internet addresses. The bur#eonin# demand for always-on!
broadband services (e.#., 14= and cable modem services) and the e/pected proliferation of wireless phones,
wireless data devices (e.#., P1,s), and eventually wireless video services may further deplete the available
IPv$ address space. .urther, if consumers are drawn to devices that can be remotely accessed and
controlled via the Internet and that re2uire fi/ed, #lobally accessible Internet addresses, demand for IP
addresses may overwhelm the remainin# pool of IPv$ addresses.
Improving Address Allocation
,doption of IPv0 could provide an opportunity to reform and rationalize the current system for allocatin#
Internet addresses, because deployment of IPv0 has created a vast new and unpopulated address space.
The historical allocation of IPv$ addresses has provided or#anizations in 9orth ,merica, -urope, and
,ustralia with the ma6ority of currently assi#ned IPv$ address bloc+s. , lar#e portion of those addresses
remain unused. ,lthou#h current allocation policies have improved, no incentives have been created to
prevent warehousin#! of IP addresses
[ 5 !
or to motivate the return of unused IP addresses. ,s a result,
many or#anizations still have very lar#e address bloc+s that have never been fully used and may never be
reclaimed in the absence of concerted action by #overnments or by Internet re#istries
1eployment of IPv0 creates an opportunity to use the lessons learned from the past to adopt more efficient
allocation policies for IPv0 addresses. ,lthou#h concerns about IPv$ address e/haustion drove development
of IPv0, steps have been ta+en to conserve addresses and to improve the efficiency of address allocation. ,s
a result, a number of observers believe that the >nited 4tates, (estern -urope, and ,ustralia may not
e/perience address space concerns for some time.
2.1.1.2 Facilitating End-to-End Services and Applications
Proponents of IPv0 contend that the massive increase in IP addresses afforded by IPv0 deployment could
stimulate development of innovative end-to-end (-%-) applications by eliminatin# the need for networ+
address translation (9,T) e2uipment. , 9,T is a device often placed between a private networ+ and the
Internet to allow a lar#e number of hosts on the private networ+ to share a smaller number of #lobally
routable, public! IP addresses for communications over the Internet.
>se of IPv0 obviates the need for 9,Ts, adoption of IPv0 could stimulate the development and deployment
of innovative -%- applications. This may occur because applications desi#ners would be able to focus on
core products and services, rather than networ+ lo#istics.
(ithout 9,Ts, moreover, applications such as ?oice-over IP (?oIP) and real-time videoconferencin# could be
implemented much more simply, because a direct connection (i.e., IP address to IP address) could be
initiated to any host, without the need to establish additional protocols and procedures to traverse one or
more 9,T devices.
Si"#lifie$ %o&ilit'
8obile services and mobile users could be ma6or beneficiaries of the massive address space available via
IPv0. 4print su##ests that the emer#ence of mobile data services such as wireless data, picture mail, and
te/t messa#in# could drive the adoption of IPv0. ,ccordin# to 8icrosoft, IPv0 better handles mobile
applications and services.! The 9orth ,merican IPv0 Tas+ .orce su##ests that IPv0 allows devices to attach
to networ+s at different points more easily than is currently achievable usin# IPv$ alternatives, principally
throu#h the use of stateless address autoconfi#uration and nei#hbor discovery capabilities. The possibility of
continuous Internet connectivity for laptops, mobile phones, P1,s, sensors, and other mobile or portable
devices, in turn, could spur development of myriad new applications in both the public and private sectors.
I"#(ove$ )*alit' of Se(vi+e ,)oS-
Internet transmission currently is a best effort! schemeusers cannot e/pect that hi#h priority! traffic will
be handled any differently from other traffic. The IPv0 pac+et header contains a fieldthe flow label!that
is not found in IPv$ and that is intended to assist with @o4. The flow label allows a user or provider to
identify those traffic flows for which the provider re2uests special handlin# by networ+ routers with #reater
specificity (or #ranularity!) than is available under IPv$
,nother constraint on the wide scale implementation of @o4, either in IPv0 or IPv$, would be the lac+ of
@o4 support in any one networ+ se#ment of the transmission path. 4uch a deficiency could ne#ate @o4
#ains realized in the rest of the networ+ path. .urther, from a commercial standpoint, service providers will
not offer @o4 support unless the offered differential in service 2uality translates into increased revenues
from customers.
Reduced Network Administration Costs
-/perts have su##ested that IPv0 will reduce networ+ administration costs in the lon# run if enterprises
reor#anize their networ+in# structure and operatin# processes to ta+e advanta#e of IPv0's capabilities and
remove 9,Ts from their networ+s. The vast number of addresses available under IPv0 could simplify (and
thus reduce the costs of) subnet mana#ement because each subnet could be #iven substantially more
address space than the number of nodes that could be connected to it. If adoption of IPv0 motivates an
or#anization to dispense with 9,Ts, networ+ administrators could more effectively use pin#, traceroute, and
other tools to dia#nose networ+ problems or to debu# applications between pairs of hosts.
Increased Overall Network Efficiency
Aemovin# 9,Ts, firewalls, and middlebo/es, andBor restructurin# networ+ routin# mechanisms (and
administrative activities) would li+ely result in fewer processin# steps and reduced transmission bottlenec+s.
The chan#e to a fi/ed header size in IPv0 could yield processin# efficiencies, and deployment of IPv0 could
also allow routin# tables to be reduced in size and redesi#ned for ma/imum efficiency.
... Sta/e0ol$e( Costs of A$o#tin1 IPv6
The potential costs associated with deployin# IPv0 consist of a mi/ture of hardware, software, labor, and
miscellaneous costs. -ach or#anization or user throu#hout the Internet supply chain will incur some costs to
transition to IPv0, primarily in the form of labor and capital e/penditures re2uired to inte#rate IPv0
capabilities into e/istin# networ+s. -/penditures and support activities will vary #reatly across and within
sta+eholder #roups dependin# on their e/istin# infrastructure and IPv0-related needs. 7y and lar#e, I4Ps
offerin# service to lar#e #roups of customers will li+ely incur the lar#est transition costs per or#anization,
while independent users will bear little, if any, costs.
Ta&le .2.. Ove(vie3 of Relative IPv6 Costs
Stake-
holders
Relative
Cost
Transition Cost
Breakdown
Timin Issues
!ey "actors in
Bearin Costs
Card-
(are
(C()
4oft-
ware
(4() =abor
Cardware
?endors
=ow )&D )&D ;&D 3urrently most are
providin# IPv0
capabilities
Aollin# in IPv0 as standard AE1
e/pense" international interest
and future profits incentives
investments
4oftware
?endors
=ow B
8edium
)&D )&D ;&D 3urrently some are
providin# IPv0
capabilities
Interoperability issues could
increase costs
Internet
>sers (lar#e)
8edium )&D %&D F&D ?ery few currently usin#
IPv0" C( and 4( will
become capable as
routine up#rade"
enablin# cost should
decrease over time
>sers will wait for si#nificantly
lower enablement costs or
(more probably) a +iller
application re2uirin# IPv0 for
end-to-end functionality before
enablin#
Internet
>sers (small)
=ow :&D $&D :&D ,vailability and
adoption schedules
(ith little money to spare,
these users must see a clear
return on investment (AGI)
Internet
4ervice
Providers
(I4Ps)
Ci#h )5D )5D F&D ?ery few offerin# IPv0
service" no demand
currently" very hi#h
cost currently to
up#rade ma6or
capabilities
I4Ps see low or none/istent
AGI, hi#h costs, and hi#h ris+