See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/283503724

A Survey of Transition Mechanisms from IPv4 to IPv6 – Simulated Test Bed

and Analysis

Conference Paper · February 2015

CITATIONS READS

0 50

3 authors:

Saadullah Kalwar Nafeesa Bohra

University of Wollongong Mehran University of Engineering and Technology
7 PUBLICATIONS   13 CITATIONS    15 PUBLICATIONS   13 CITATIONS   

SEE PROFILE SEE PROFILE

Aftab Ahmed Memon

SE, Canada
58 PUBLICATIONS   142 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

LEO-MESH Nets View project

Light Fidelity View project

All content following this page was uploaded by Saadullah Kalwar on 13 December 2017.

The user has requested enhancement of the downloaded file.

 A Survey of Transition Mechanisms from IPv4 to
IPv6 – Simulated Test Bed and Analysis
Saadullah Kalwar, Nafeesa Bohra, Aftab A. Memon
Department of Telecommunication Engineering, Mehran University of Engineering and Technology, Jamshoro, Pakistan
Email: [saad.kalwar, nafeesa.zaki,aftab.memon]@faculty.muet.edu.pk

Abstract— The introduction of IPv6 has opened up several which increases the total number of IP nodes from 232 addresses
questions with reference to its adaption and transition from IPv4 of class full addressing and somehow delayed the transition
to IPv6 and is one of the crucial issues being frequently discussed process [2]. The objectives of this survey paper are twofold.
in networking community today. IPv6 provides many seamless Firstly, the paper highlight the issues related with the transition
features that makes it far better protocol as compared to its
from IPv4 to IPv6. Secondly, the aim is to look into the
predecessor IPv4. It is a well-known fact that IPv4 is a defacto
standard at present and is currently been deployed in almost all transition mechanism that can be provided seamlessly to end
the Internet architecture, hence the transition process from IPv4 users where they will be able to use all the services already
to IPv6 is very challenging. In order to avoid the transition, or in being used over IPv4.In order to achieve the said objectives a
actual sense to delay it, many techniques have been introduced simulated test bed has been deployed at Mehran University of
such as CIDR and NAT but the fact is, the pool of IP addresses is Engineering and Technology (MUET), Jamshoro, Pakistan.
depleting and ultimate solution is to move towards IPv6. The The purpose is to observe and tackle the issues and challenges
objectives of this survey paper are twofold. Firstly, to highlight the that are likely to be faced during the transition from IPv4 to
issues related with the transition from IPv4 to IPv6. Secondly, to IPv6. Graphical Network Simulator 3 (GNS3) and Wireshark
find the transition mechanism that can be provided seamlessly to
are used for simulation and Dual Stack Transition Mechanism
end users where they will be able to use all the services of IPv4. In
order to achieve the said objectives a simulated test bed has been (DSTM) has been chosen for the test bed. DSTM allow both
deployed at Mehran University of Engineering and Technology protocols to run simultaneously and the results show that it
(MUET), Jamshoro, Pakistan. The purpose is to tackle the issues provide seamless transition from IPv4 to IPv6.
and challenges that are likely to be faced during the transition The rest of paper is organized as: Section 2 gives the overview
from IPv4 to IPv6. GNS3 and Wireshark are used for simulation and motivation, and mainly discusses the constraints that have
and DSTM has been chosen as the transition mechanism for the been responsible for the delay of transition and need for the
test bed. DSTM allow both protocols to run simultaneously and transition. The comparison between NAT and IPv6 is provided
the results show that it also provide seamless transition from IPv4 to show difference between both the choices. Section 3
to IPv6.
introduces the transition strategies. Section 4 presents the
Keywords—IPv6; transition strategies; NAT; dual stack; simulated test bed and discusses why DSTM have been chosen
tunneling for the test bed. Finally, Section 5 conclude the paper and give
future directions.
I. INTRODUCTION
In 1970s when Internet started to evolve from ARPANET,
intentions about Internet were not something what we see
today. The Internet was primarily introduced with the intention
to connect few agencies of US but later on it evolved into a
network of networks connecting complete globe. The current
Internet is based upon Internet Protocol ver4 (IPv4) which is
inherently light and simple [1]. With the increasing use of IP, it
started to become over whelmed because of lower features of
security etc. In order to cater the demands, a lot of extensions
were introduced such as Classless Inter Domain Routing
(CIDR) [2], Address Resolution Protocol (ARP) [3], Dynamic
Host Configuration Protocol (DHCP) [4] and Network Address
Translation (NAT) [5]. With lot of extensions, the working of
simple IPv4 over public networks became very complicated. It
gets even worst when the Internet users started to grow in
geometric progression. This growth ultimately resulted in
shortage of IP addresses because IPv4 supports only 32-bit
address which means 4.3 billion addresses, whereas addresses
based on classes would only be less than a Billion. To cater
Fig. 1. Working of NAT
these problems, CIDR and NAT mechanisms were introduced

ISBN: 978-1-4799-6376-8/15/$31.00 ©2015 IEEE 30

 II. OVERVIEW AND MOTIVATION
There are many constraints such as: whether IPv6 would be
adapted by the networking community, how is the transition
going to take place, have been delaying the transition process.
Although, the delay is not a long term solution when one look at
the evolution of current Internet. The ultimate solution is
transition to IPv6.
Most widely discussed topic in networking community today
is either to adapt IPv6 or to stay with IPv4? If both of them are
compared, the advantages offered by IPv6 can clearly be seen
over its predecessor (IPv4), hence the question arises what holds
the transition? First answer to this question is that the two
protocols are not directly compatible to one another. Completely
new infrastructure is required in order to make them compatible
which is neither preferred nor recommended. Therefore, it is
recommended to keep running the existing protocol as long as
possible. If IPv4 is used than the question arises, IPv4 is being
used because of NAT? NAT makes the network a private
network by making use of private IP addresses which are not
routable, rather it is connected to the router responsible for the
translation of private IP into public IP address [5]. In this way
by using a single public IP, set of stations can be connected to
Internet, and the private IP addresses being used by them would
also be used by other private networks, as shown in Fig 1.
The mechanism shown in Figure 1, illustrate that there is an
increase in the total number of IP addresses, hence, no need to
provide unique IP to each and every host over the Internet. Part
of network shown inside a rectangular box (Figure 1) uses
private IP addresses and have introduced one more discussion
that either to go for IPv6 or keep using NAT. The comparison
between IPv6 and NAT is given in Table. I
TABLE I: COMPARISON BETWEEN IPV6 AND NAT
the total population across the globe and will also provide
sufficient addresses for the generations to come.
Keeping all these things in view, network of MUET is taken
into account in this research and complete transition layout have
been proposed. Before discussing the transition layout,
transition strategies are discussed in the Section 3.
III. INTRODUCTION TO TRANSITION STRATEGIES
IPv4 and IPv4 are not directly compatible with one another.
Therefore it is important to develop a mechanism so that smooth
transition is possible which enables the applications to continue
working while the network is being upgraded [7]. In order to
perform transition from IPv4 to IPv6 in a transparent manner
transition strategies have also been evolved along with the
evolution of IPv6. Transition strategies are categorized as
Translation Mechanisms, Tunneling and DSTM [8], and are
briefly discussed one by one.
A. Translation Mechanism
Translation Mechanism is meant for communication
between IPv4 and IPv6 network. Basic mechanism behind the
strategy is header translation due to which it is known as
translation mechanism. It can be considered similar to NAT.
NAT translates between private and public IP addresses, here
IPv4 and IPv6 headers are translated to each other [9]. For
example a packet originating from IPv4 network, the translator
would convert its header into IPv6 header before it is sent to
IPv6 network and same process is done in inverse manner too,
and is illustrated in Fig. 2. In this technique, separate translator

NAT
NAT doesn’t provide end
to end connectivity to
hosts
NAT is not a long term
solution
NAT provides isolation
benefit of security
IPv6
IPv6 global address provides
end to end connectivity
IPv6 is the actual solution due
to theoretically unlimited
address space
Less secure due to direct
connectivity
Fig. 2. Translation as Transition Mechanism
is required between both the networks, programming translators
is difficult task and sometimes high capacity translators may be
required. It faces similar security issues as are faced in NAT
because there is no end to end connectivity. On the other hand,
it can be useful in some scenarios such as if it is required to
connect the IPv6 and IPv4 nodes independently then Translation
Mechanism is the best choice [10].

From Table I, it can be observed that NAT is not a solution B. Tunneling

to the address shortage problem, it is just a way to keep this
problem on hold and keep using Internet, actual solution is IPv6.
IPv6 eliminates the need of NAT by providing huge number of
addresses such that every node on the Internet may have unique
IP address because it is inherently 64-bit and there will be no
need of NAT [6]. Transition from IPv4 to IPv6 is also the need
of time because of the incoming boom of cellular generation,
where every mobile phone will be a unique IP node. The
transition is also required because the way the world population
is increasing, more IP addresses are needed to facilitate the
mobile users. Hence, the transition towards IPv6 is the only
solution which gives the total number of addresses lot more than
Tunneling is very different as compared to translation. It is
used to connect two IPv6 nodes using IPv4 network. Hence,
based on current scenarios where IPv4 is still dominant, if two
IPv6 networks would communicate with each other, the feasible
scheme is tunneling. Tunneling encapsulates packets of one
protocol onto another. Packets originating from IPv6 node
would be encapsulated within IPv4 packets, in this way they
propagate along IPv4 network [11], as shown in Fig. 3.
C. Dual Stack Transition Mechanism
The most widely used transition mechanism across the world is
DSTM [12]. Mainly because of the ease of implementation and
support for both (IPv4/IPv6) protocols. In order to migrate to

ISBN: 978-1-4799-6376-8/15/$31.00 ©2015 IEEE 31

IPv6, IPv4 cannot be completely removed because of the reason
that the current infrastructure is based on IPv4 and needed to be
replaced gradually as IPv6 will grow. During the transition
unchanged. In case of PC having Windows XP, IPv6 can be
enabled easily using ipv6 enable in command prompt [14]. The
reason behind using the Cisco routers and switches is that Cisco
Software Release 12.0S enabled the support of IPv6 in core
service provider infrastructures. Currently Cisco IOS releases

Fig. 5. Test Bed Setup

enable IPv6 in a wide range of Cisco products. Simulation test

bed have been developed using Wireshark together with GNS3,
Fig. 3. Tunneling as Transition Mechanism
that allows emulation of complex networks because It allows to
use Cisco IOS in virtual environment [15]. Simulation test bed
is shown in Fig 5. Fig 5 is the dual stacked network i-e every
node has been assigned IPv4 address as well as IPv6 address.

TABLE II: DUAL STACK IP ADDRESSES

Router Interface IPv4 Address IPv6 Address

MUET e0/0 192.168.0.2/30 2001:DB8::3/64

MUET e0/1 172.16.23.3/24 2001:DB8:5::1/64

Fig. 4. Dual Stack as Transition Mechanism
Dist: f0/0 192.168.0.6/30 2001:DB8:1::4/64
phase, both protocols would be used; DSTM offers this service
[13]. DSTM allows that both IPv4 and IPv6 addresses are Dist: f1/0 12.0.0.1/24 2001:DB8:3::3/64
assigned to every node on the network and in this way it is
Dist: f1/1 192.168.2.1/24 2001:DB8:4::3/64
possible to utilize both the services at the same time as shown
in Fig. 4. Dist:2 f0/0 192.168.0.10/24 2001:DB8:2::4/64
Core e0/0 192.168.0.1/30 2001:DB8::4/64
IV. SIMULATION TEST BED
Core f1/0 192.168.0.5/30 2001:DB8:1::3/64
For analysis a simulated test bed was created based on the
network of MUET, Jamshoro Pakistan. Since it is difficult to Core f2/0 192.168.0.9/30 2001:DB8:2::3/64
take into account the complete network of MUET, Jamshoro, 2001:DB8:3:0:A978:69B3
Host 12.0.0.2/24
Pakistan, therefore, standard Cisco hierarchal network is 8D6F:2F44
considered for simulation. The network is based on Cisco C3 DHCP 2001:DB8:5::1/64
standard layered hierarchy consisting of core, distribution and
access layer. In order to carry the simulation, Cisco 3645 router
have been used comprises of IOS 12.4 and is compatible with Tests show that both protocols would be working over the
IPv6. The edge router of MUET is Cisco 3845, whereas Cisco network without interfering each other. IOS image used for
cat 4006 switch being used at core layer comprises of IOS 12.4 simulation is c3640-jk9o3s-mz.124-16.bin with IOS version
and 12.2 respectively. Similarly at distribution layer, the 12.4. Addressing scheme is shown in Table II. To see how both
network of MUET has Cisco 3550 and 3750 switches, both have protocols coexists, Wireshark is used. Wireshark is a packet
IOS version 12.2. Switches at access layer are Cisco 2950, and sniffing software which is used as network protocol analyzer. In
have IOS version 12.1. At the user end, everything will remain order to show the packets of both protocols moving over the

ISBN: 978-1-4799-6376-8/15/$31.00 ©2015 IEEE 32

same link, packets have been captured over a MUET dist switch
as shown in Fig. 05. Hello packets of EIGRP and OSPF can be
seen in Fig 6. EIGRP is running over IPv4, its multicast address
Fig. 6. Wireshark Hello Packets
Fig. 7. Ping over IPv4
Fig. 8. Ping over IPv6
is 224.0.0.10 which can be seen, similar address for OSPF in
IPv6 is ff02::5 which can also be seen. To check connectivity,
ping MUET from dist switch, the results shown in Fig 7 have
been achieved.
Fig 7 shows ping requests and replies over IPv4, we now
check connectivity from same routers over IPv6. Fig 8 shows
connectivity over IPv6. The protocol used for ping in IPv6 is
ICMPv6 similar to ICMP in IPv4. The connectivity between
both (IPv4/IPv6) the protocols has been discussed in the above
Sections, shows that for a seamless transition, it is necessary that
end user should not be affected which is only be possible when
user gets all the services over new protocol (IPv6) which were
available over IPv4. In order to test that, some basic services
were run over the network. To show how users will use
applications, IPv6 Internet is required which is not yet available
in Pakistan, hence, we made our own servers which are shown
in test bed (Fig 5) by connecting to cloud C3. These servers were
made on separate machines and linked up to GNS3 through
physical interface. Hence, works like our own Internet cloud. To
simulate the end user, VMware virtual machine has been used
as host which is connected to ICT switch in the test bed. Virtual
machine has been linked with GNS3 with the help of loopback
adapter. Windows XP operating system is used for virtual
ISBN: 978-1-4799-6376-8/15/$31.00 ©2015 IEEE
machine. Windows XP (and all newer operating systems of
Microsoft) has built in support of IPv6. But IPv6 is disabled in
Windows XP by default. We can simply enable it by entering
IPv6 enable in command prompt. When http server is accessed
from the virtual machine, result shown in Fig 9 are achieved.
Fig. 9. HTTP server being accessed over IPv6
Figure 9 shows end to end connectivity as well as it also shows
how the transition from IPv4 to IPv6 is provided without
affecting the end user applications.
As explained, DSTM fulfills all the requirement needed for
transition from IPv4 and IPv6, whereas other transition
mechanisms (Translation and tunneling mechanism) do not
allow the use of both the protocols simultaneously. Most of the
equipment and software today already support DSTM. Most of
the major websites throughout the world are already dual
stacked. Keeping all this in mind, DSTM has been chosen as the
transition mechanism for the test bed, this will allow both
protocols to run simultaneously and provide seamless transition
from IPv4 to IPv6. The addressing scheme used for the test bed
under consideration is explained in the following subsection.
A. Addressing Scheme
Current subnet structure of MUET is 172.16.x.0/24, where
x is VLAN number. Great address space of IPv6 again provides
advantage here. More number of VLANs can be created and
more hosts can be configured for each VLAN. Creating VLANs
in IPv6 is similar to that of IPv4. We can use link local
addresses. A possible structure of VLANs can be:
FF00:0:0:x::/64, where x is the VLAN number. So with this
addressing scheme, we can assign similar subnet numbers to all
VLANs and well still be having a lot of free subnets available
for future use. For example core router has IPv4 address
172.16.60.0, so its IPv6 address will be FF00:0:0:60:: VLANs
will be created in the same manner. Here example of ICT
building can be taken. The VLAN number of ICT is 23 so the
IP addresses assigned belong to 172.16.23.0. IPv6 addresses for
the hosts of ICT will be in network FF00:0:0:23::. A host
having IPv4 address 172.16.23.10 will have IPv6 address
FF00:0:0:23::10 [16].
V. CONCLUSION
In this survey paper, the issues and challenges related
with the transition from IPv4 to IPv6 have been analyzed. After
looking at different transition strategies (Translation, Tunneling
33
and DSTM) it is concluded that DSTM provide seamless [4] R. Droms, J. Bound, B. Volz, T. Lemon, C. Perkins, and M. Carney,
“Dynamic host configuration protocol for ipv6 (dhcpv6),” RFC 3315,
transition from IPv4 to IPv6. Furthermore the major issues Tech. Rep., July 2003.
faced during the transition process are also discussed. Actual [5] D. Wing, “Network address translation: Extending the internet address
test bed of MUET has been used for real time simulations and space,” IEEE Internet Computing, vol. 14, pp. 66–70, July/August 2010.
the parameters used are practically applicable to any network [6] M. Boucadair, J.-L. Grimault, P. Levis, A. Villefranque, and P. Morand,
for transition from IPv4 to IPv6. Through analysis it has been “Anticipate ipv4 address exhaustion: A critical challenge for internet
proved that DSTM is the most appropriate option and seamless survival,” Evolving Internet, International Conference on, vol. 0, pp. 27–
32, 2009.
transition from IPv4 to IPv6 is provided without affecting the
[7] M. E. Fiuczynski, V. K. Lam, and B. N. Bershad, “The design and
end users applications. It is further concluded that, most of the implementation of an ipv6/ipv4 network address and protocol translator,”
hardware and software are also supporting the IPv6 and the in Proceedings of the Annual Conference on USENIX Annual Technical
decision of transition strategy depends upon the type of the Conference, ser. ATEC ’98. Berkeley, CA, USA: USENIX Association,
network. Transition from IPv4 to IPv6 is a planned process. 1998, pp. 17–17. [Online]. Available:
http://dl.acm.org/citation.cfm?id=1268256.1268273
Major players all over the world have already started the
[8] I. Raicu and S. Zeadally, “Evaluating ipv4 to ipv6 transition
process. It is the high time to move towards IPv6, before the mechanisms,” in Telecommunications, 2003. ICT 2003. 10th
acute need arises. International Conference on, vol. 2, Feb 2003, pp. 1091–1098 vol.2.
[9] M. B. M. B. X. L. [6] C. Bao, C. Huitema, “Ipv6 addressing of ipv4/ipv6
ACKNOWLEDGMENT translators,” October 2010, rFC 6052.
We would like to acknowledge staff of ICPC, Mehran [10] CISCO, “NAT64 Technology: Connecting IPv6 and IPv4 Networks,”
White paper, Tech. Rep., 2012.
University of Engg. and Tech. Jamshoro, Pakistan for his
[11] S. D. A. Conta, “Generic Packet Tunneling in IPv6,” Cisco Systems,
valuable guidance throughout this research work. Without their Tech. Rep., 1998.
kind support this work would not have been possible. [12] X. D. Lewis, “Ipv6: Current deployment and migration status,” in
International journal of research and reviewsin Computer Sceince
REFERENCES (IJRRCS), june 2010.
[13] M. Dunmore, “An ipv6 deployment guide, 6net,” in The 6NAT
[1] R. Marina, “Rfc 791 internet protocol - darpa inernet programm, protocol Consortium, September 2005.
specification,” 1981. [Online]. Available: http://tools.ietf.org/html/rfc791 [14] “Cisco router guide,” summer, vol. 6, 2010.
[2] V. Fuller, T. Li, J. Yu, and K. Varadhan, “Classless Inter-Domain Routing [15] M. Fuszner, “Graphical network simulator.” [Online]. Available:
(CIDR): an Address Assignment and Aggregation Strategy,” RFC 1519 http://www.gns3.net/
(Proposed Standard), Internet Engineering Task Force, September 1993, [16] T. Chown, “Use of VLANs for IPv4-IPv6 Coexistence in Enterprise
obsoleted by RFC 4632. [Online]. Available: Networks,” RFC 4554 (Informational), Internet Engineering Task Force,
http://www.ietf.org/rfc/rfc1519.txt June 2006. [Online]. Available: http://www.ietf.org/rfc/rfc4554.txt
[3] M. G. Gouda and C.-T. Huang, “A secure address resolution protocol,”
Computer Networks, vol. 41, no. 1, pp. 57–71, 2003.

ISBN: 978-1-4799-6376-8/15/$31.00 ©2015 IEEE 34

View publication stats