Int. J. Communications, Network and System Sciences, 2015, 8, 62-69
Published Online April 2015 in SciRes. http://www.scirp.org/journal/ijcns
How to cite this paper: Mi, W. and Zhang, X.D. (2015) The Applicability Analysis of IPv6 Translation Transition Mechanisms.
Int. J. Communications, Network and System Sciences, 8, 62-69. h ttp://dx.d oi.org/10.4236/ ijcns.2015.84008
The Applicability Analysis of IPv6
Translation Transition Mechanisms
Wei Mi, Xiaodan Zhang
Computer Network Information Center, Chinese Academy of Sciences, Beijing, China
Email: email@example.com, firstname.lastname@example.org
Received March 2015
Due to the exhaustion of IPv4 address resources, the transition from IPv4 to IPv6 is inevi table and
fairly urgent. Numerous tra n si tion mechanisms have been proposed to solve challenging issues of
IPv6 transition. An inter-c on n ec ti on betwee n IPv4 and IPv6 networks or hosts requirement has
been happening throughout the IPv6 tran si t i on process. And one -ti me translation scheme is in-
dispensable to achieve the in ter -connecti on. In addition, double translation can be used in the
IPv4-IPv6-IPv4 s cenario. As a long-term strategy, tran s lat i on scheme is important an d ine vit ab le.
However, becaus e of the diverse characteristics and transition requirements of practical networks
and the lack of applicability an aly sis, th e selection and deploymen t of transition mechanisms are
facing with gran d challenges. Targeting at those challenge s, th i s paper investi ga tes th e basic issues
and key elements of IPv6 translation transition mechanisms, and presents its first appli cability
index system. In particular, we analy ze the applicab ility of existing proposed translation tech-
niques based on the pre sent ed in dex system, which has significant guidance in the practical dep-
loyment of IPv6 transition techniq ues.
IPv6 Transition, Translation Mechanisms, Applicability, Index System
With the rapid growth of Internet scale, the exhaustion of IPv4 addresses is a significant problem. IPv6 was de-
signed to be an evolutionary step from IPv4, overcoming the problems of IPv4 and promoting the development
of Next Generatio n In t ern e t. Due to th e incompatibilit y in nature, IPv 6 transitio n will face many techn ical chal-
lenges, such as heterogeneous addressing, different semantic, routing isolation, huge size and the transparent to
users and to app lications . I Pv6 transition is a world recognized significant technology problem in the develop-
ment of Nex t Ge neration Internet.
During the process of IPv6 trans ition, no matter whic h network protocol is used, it must support both I Pv4
services and IP v6 s ervices, and ensure the transparence to the upper layer applications. The character istics and
transition requirements of practical networks are diverse and complex. Theoretically, the IPv6 transition scena-
rios can be decomposed into two types: inter-connection (IPv4-IPv6) and heterogeneous traversing (IPv6-IPv4-
IPv6 or IPv4-IPv6-IPv4). An inter-connection happens when networks or hosts using different address families
W. Mi, X. D. Zhang
are directly connected and want to communicate. A traversing happens when two or more native IPv4/IPv6
networks (or hosts) are separated by a network which uses the other address family and thereby is not IPv4/
Numerous transition mechanisms have been proposed to solve challenging issues of IPv6 transition, which
can be divided into dual-stack, tunnel, and translation me chanisms. Dual -stack mechanisms can support both
IPv4 and IPv 6, bu t they bring the high cost on both the hardware upgrading and ne twork o peration/management.
Tunnel mechanisms possess the merits of expansibility, flexibility and simplified achievement. And the Softwire
Working Grou p of IETF is focusing on developing and standardizing the tunnel mechanisms. However, tunnel
scheme cannot resolve the inter-connection problem which has been happening throughout the IPv6 transition
process. Compared with dual-stack and tunnel, translation can be used to achieve direct communication between
IPv4 and IPv6. After development these years, translation technology has already formed its own one more
complete system which had been developed and standardized in the Behave Working Group of IETF. As a
long-term strategy, trans lation scheme is important and inevitable.
In practical deployment, it is important to find feasible trans ition mechanisms and make approp ria te plan to
cover all potential communication scenarios. Howeve r, it brings great challenges to the research community of
IPv6 transition. W ith the diverse characteristics and transition requirements of practical networks and the lack of
overall transition architecture, the selection and deployment of IPv6 transition mechanisms are very difficult.
Thus, there is a strong need to take the research on the applicability of transition mechanisms. However, the ap-
plicability criterion and analysis are lack in the current literature review.
In an effort to push forward the IPv6 transition process, this paper deeply discusses the basic issues and key
elements of IPv6 translation mechanisms, and presents the first applicability index system and analyzes the ap-
plicability of existing translation techn ique s. All of th e se have guiding significantly in the IPv6 tran sition
The rest of this paper is organized as follows. Section 2 presents the related work. The applicability index
system and the applicability analysis of mainstream IPv6 translation mechanisms are shown in Section 3 and 4.
Finally, Section 5 concludes this study.
2. Related Work
Numerous stud ies on the evaluation of IPv6 transition mechanisms have been reported in the current literature.
Shin et al.  showed the impact of IPv6 transition mechanisms on user applications. Law et al. in  focu sed
on the performance of dual-stack technologies in terms of various network metrics including network connec-
tivity, hop-count, RTT, throug hpu t, operating systems dependencies and the address configuration latency. The
authors in   provided the evaluation of tunnel mechanisms with the key performance-related metrics in-
cluding throughput, delay, jitter, and the CPU usage of transition nodes. AlJa’afreh et al.  gave the compa-
rison between the tunneling process and mapping schemes for IPv4/IPv6 transition using en d -to-end delay and
throughput as the key performance metrics. Guer i n and Hosanagar  adopted a simple model to illustrate how
the connectivity quality affects both IPv6 adoption and the volume of translation traffic, and summarize their
implications for IPv6 adoption. The authors in  evaluated the dual-stack protocol and tunneling transition
based on the metrics of throughput and round-trip delay. Several studies - presented the comparisons of
translation mechanisms with the aspects of operation complexity and scalability, real-time communications,
field device, multicast address, and application-layer protocol. The authors in  proposed the evaluation of the
transition mechanisms including the estimation on the scalability, heterogeneous addressing and application-
layer translation, hardware cost, performance and capacity of the equipment, security, end-to-end property, and
the influence for developing applications.
Compared with existing studies, in , we also provided unified assessment criterion in terms of functiona-
lity, applications, performance, development and secur ity to evaluate the mainstream transition mechanisms.
The unified evaluation criter ion is sh o w n in Table 1.
3. The Applicability Index System
3.1. Basic Problems
Translation scheme is used to achieve direct communication between IPv4 and IPv6 networks (or hosts). Its ba-
sic operation is to co n vert the s emantics between IPv4 and IPv6, turning IPv4 packet into IPv 6 if the packet is
W. Mi, X. D. Zhang
Table 1. The evaluation criterion of IPv6 transition mechanisms.
Evaluation criterion Description
Functionality Including the transition scenario, transition function, equipment requirements,
and IPv4 and IPv6 address requirements.
Applications The impact on IPv4 or IPv6 application.
Performance The performance evaluation is mainly for the equipments, such as forwarding performance,
and the searching, storage, and computational overhead.
Deployment Including the implementation cost, and the ease of management and usage.
Security Including the security issues and concerns.
destined to IPv6 network, or turning IPv6 packet into IPv4 if th e packet is destin ed to IPv4 network.
The basic data operation
IPv4-IPv6 packet translation is the basic data plane operation. It involves network, transport, and application
layer. Thus, the basic operations include addre ss and port conversion, IP/TCP/UDP protocol field translation,
and application layer translation (addr es s and po r t conversion when they appe ar in application protocol). What is
more, to overcome further diversities in the protocol definition between IPv4 and IPv6, translation has to take
care of issues like fragmentation and reassembling, path MTU (Maximum Transmission Unit) discov er y, ICMP,
The basic control operation
The basic control plane operation is the address conversion rule: either some special address scheme needs to
be deployed in advance, or dynamic address bindings have to be built during the translation. Heterogeneous ad-
dressing (learning the in-protocol address of the remote end) and corr espond ing routing should be performed
based on the address conver sio n ru le .
The translation model
According to the applicable scenarios, the models of IPv6 translation mechanisms are divided into ap p lica-
tion-side, network -side and ho s t-side translation. Accor ding to the address conversion mann er, we can a l so di-
vide network-side tr anslatio n mechanisms into stateless tr anslatio n and stateful trans lation.
3.2. Key Elements
In translation technologies, the translator is the transitio n equipments. Usually, network-side translation hap-
pens on the IPv4-IPv6 border, so the tr anslator would be an AFBR (Address Family Border Router). And
host-side translation happens in the TCP/IP stack of the end host, so the transla tor would be host. They s hould
support address and port conversion, IP/TCP/UDP protocol field translation and also maintain the state. Thus,
translator has requirements in the use of bandwidth, computing and finding, storage.
Address tr anslation
Using specific addres s conver sion ru le s , the translator either gets I Pv4 a ddre ss fr o m a specific position of
IPv6 address, or builds the IPv6 addr ess using IPv4 address . The address conversion rule includes some special
address sch eme needs to be deployed in advance, and dynamic address bindings have to be built during the
Other fields translation
Except the source address and destination address, there are other fields (fragmentation ID, checksum, ban-
fragmentation flag, etc.) in IP/ICMP packets, wh ic h ca rry sp ecif ic information. In the IPv4-IPv6 translation
process, the header information of IPv4 packet cannot be lost. Therefore, translation techniques are required to
ensure the integrity of IPv4 inf ormatio n .
In IPv6 transition, the state is available to only represent a series of attributes mapping relationship among a
communication entity (such as client, server), a communication path (the connection between communication
entities), or a communication process (control flow an d data flow). The IP address in network layer and ID in
transport layer of the communication entities are the core state which need translation techn iqu es to maintain.
W. Mi, X. D. Zhang
DNS64 and DNS46
The main function of DNS64 and DNS46 is to realize the bi-directional translation between A and AAAA
record. DN S6 4 translates the AAAA query from IPv6 hosts into A query when r eceivin g one, and DNS46
translates the A response for IPv6 hosts into AAAA response following the IPv4-mapped address rul e before
sending one out. The re are usually two kinds of implementations: static configu ra tion DNS records and dynamic
Application layer translation
Those applicatio ns whose addres s and port co nv ersion w hen they appear in application protocol cannot work
in NAT and IPv4/IPv6 translation environments. Such as FTP, SIP, etc. One solution is assist applications to
realize translation work by using the application layer gateway.
3.3. Applicability Index System
Based on the analysis of basic problems and key elements, we built the first applicability index system in ter ms
of sustainable, applications, performance and development to evaluate all IPv6 translation transition mechan-
isms. The applicability index system is shown in Table 2.
4. The Applicability Analysis
In the all stages of transition from IPv4 to IPv6, IPv4 networks/hosts and IPv6 networks/hosts are likely to
coexist. For the operators, supporting inter-connection is inevitable. One-time tr ans lation mechanisms emerged
Table 2. The applicability index system.
Applicability criterion Description
Scenarios and function of transition Whether meet the needs of
The coupling degree between IPv4 address and IPv6 address.
The reuse rate of IPv4 addresses resource. Whether promote the deployment
and usage of IPv6.
The support degree
of business application.
The support degree of IPv4 application. End-to-end property. Impact on the IPv4 business
The support degree of IPv6 application.
Impact on the IPv6 business
The performance requirement of
The routing information
announcement The capacity of bandwidth,
computing and finding, storage.
The space and time overhead
of state maintenance
Routing scalability The aggregation of
IPv6 addresses Impact on the scope of deployment.
Robustness The capacity of redundancy backup.
The cost of development
Technological and industry maturity The support degree of IETF The support degree of standard
Impact on application layer.
The impact on the present network. Impact on network layer.
Impact on end users layer.
The cost of operation, management
It impact on operator. Maintenance
Security Including the security issues and concerns. The security of translation
W. Mi, X. D. Zhang
With the IPv6 development, IPv4 Internet has been gradual ly replaced. For the low cost, network operators
tend to bu ild IP v6 n e two rk rather than dual-stack n etwor k. In ord er to ensu re the compatibility of legacy IPv4
application, IPv4-IPv6-IPv4 double translation scheme is provided.
According to the different transition stages, this paper will analyze the applicability of one-time translation
mechanisms in the all stages and dou b le translatio n mechanisms in the middle- to-late-stage.
4.1. The Applicability Analysis of One-Time Translation Mechanisms
IETF h as developed and standardized many one-time translation transition mechanisms, such as network-side
translation mechanisms whi ch can be divided i nto stateless tra nslation (SIIT , IVI ) and stateful transl a-
tion (NAT-PT , NAT64 ), and host-side translation mechanisms (BIH ). In this section, we will
analyze the applicability of these mechanisms, which is summarized in Table 3.
SIIT is an early stateless tr a ns lation mechanism. It does not introduce new security issues to the network.
However, the usage of fixed address prefix brings significant routing scalability problem. Due to the per-host
IPv4 address consumption requirement, the IPv6 side of SIIT cannot be huge which is unfavorable to IPv6 tran-
sition. Therefore, SIIT apply to the early-stage of IPv6 transition.
IVI follows the principle of stateless translation and improves SIIT. The use of NSP (network-specific, varia-
ble prefix) makes the rou ting sca lability is no longer a concern. However, the per-host IPv4 ad dress consump-
tion is still required in IVI. Therefore, the IPv6 side cannot be huge and IVI also apply to the early-stag e of IPv6
NAT-PT is originally adopted to achieve the heterogeneous addressing procedure for both sides with DNS-
ALG on th e trans lator. However, it possesses several technical issues, e.g., the d isruption of en d -to-end transp a-
rency, information loss and misunderstanding of DNS-ALG, low scalability of network topology caused by the
deep-coupling between DNS-ALG an d translator, and the inability to redirect traffic with persistence address-
As an enhan ced mechanism, NAT64 extracts the DNS-ALG function fro m the translator and makes it become
a dedicated DNS64 server. However, NAT64 only specifies the communication initiated from the IPv6 side. It
also destroys the end-to-end transparency. NAT-PT and NAT64 have DoS attack on the binding table, with in-
gress filtering on the IPv6 s ide as th e solution.
In the host-side translation mechanisms, the translation is inside the TCP/IP stack of the host, which may re-
duce the deployment overhead of network. However, due to the scale of terminal hosts is much larger than the
device in th e network side, it will bring a great pressure to ISP and be unfavorab le to I Pv6 transition.
4.2. The Applicability Analysis of Double Translation Mechanisms
At present, IETF has been developing and standardizing double translation transition mechanisms. The main-
stream mechanisms include 464XLAT , d IV I  and MAP-T [2 0 ]. The applicability analys is of thes e me-
chanisms is provided and summarized in Table 4.
464XLAT suppor ts an IPv4 host (especially mobile host) to connect to the IPv4 Internet throug h the IPv6
network. Due to the combination of SIIT on th e user sid e and NAT64 on the carrier side, 464XLAT faces the
issues in both SIIT and NAT64.
As an extension of IVI, dIVI is a stateless net wo rk -sid e transla tion in wh ich one IPv4 address is shared by
multiple IPv6 hosts through port space division. However, dIVI cannot realize the inter-connection between two
IPv4 hosts insider the two dIVI network separately.
In MAP-T, the I Pv6 packets have to use the I Pv4-mapped addresses as source addresses in the inbou nd dir ec-
tion and destination addresses in the outbound direction. However, since the mechanism deeply couples IPv4
and IPv6 addres sin g, it becomes a little less flexible to deplo y: the deployment has to be entire-network style
rather than on demand style, otherw ise some of the coup led IP v4 addresses will be wasted.
For the consideration of deployment scenarios and addres s format, numerous translation transition mechanisms
have been proposed in IETF. However, due to a wide range of mechanisms and a lot of overla p a nd similar
W. Mi, X. D. Zhang
Table 3. The applicability analysis of one-time translation mechanisms.
Applicability SIIT IVI NAT-PT NAT64 BIH
Scenarios and function of
between IPv6 network
and IPv4 network/Internet
IPv4 network, and
from IPv6 network
to IPv4 Internet.
on the end
to an IPv6
Coupling degree between
IPv4 address and IPv6 address. High by stateless
address mapping. Medium . Low.
Reuse rate of IPv4
addresses resource. Low High by stateful address + port binding low
Support degree of IPv4
application. High Low by destroying the end-to-end
transparency Low in BIS,
high in BIA.
Support degree of IPv6
application Low High
of: ffff:0:0/96 to
IPv6 side and the
address to IPv4
the prefix of
IPv6 hosts to
IPv4 side and
the NSP route
to IPv6 side.
It should advertise
the prefix of IPv4
address pool to IPv4
side, as well as the
IPv6 prefix of:
96 to IPv6 side.
It should advertise
the prefix of 64:
FF9B: /96 to IPv6
side, and the prefix
of the IPv4 address
pool to IPv4 side.
The space and
low High. Per-flow-state maintenance. High in BIS,
low in BIA.
ggregation of IPv6 addresses)
Low by using
the fixed prefix. High by
using NSP. No impact.
Robustness High Low
The cost of
Technological and industry
maturity RFC 2765 RFC 6219 IETF had
discarded it. RFC 6146 RFC 6535
Impact on application layer. No impact. Network gateway
NAT-ALG. No impact.
on the socket
Impact on network layer. Network gateway
should support the
The scalability of
server should be
Impact on end users layer. No impact
of the host
Configuration Medium High
Maintenance Low High High
Troubleshooting High Low Low
Security The security issues and
SIIT and IVI do not
introduce new security
issues to the network.
DoS attack on the binding table.
And ingress filtering on the IPv6
side as the solution. DoS attack.
W. Mi, X. D. Zhang
Table 4. The applicability analysis of double translation mechanisms.
Applicability 464XLAT dIVI MAP-T
Scenarios and function of transition IPv4 mobile host-
Coupling degree between IPv4
address and IPv6 address. High
Reuse rate of IPv4 addresses resource. High
degree of business
Support degree of IPv4 application. High
Support degree of IPv6 application Low
Smilar with SIIT
and NAT64 Smilar with IVI No need.
The space and time
overhead of state
Low in CLAT,
high in PLAT Low Low space and
high time overhead.
(Aggregation of IPv6 addresses) Low High low
The cost of
Technological and industry maturity RFC 6877 IRTF draft (work in progress)
Impact on application layer. No impact.
Impact on network layer. Network gateway
Network gateway should
support MAP-T CE and BR
Impact on end users layer. Host equipment
CLAT No impact
Configuration High Medium
Maintenance Medium Low
Troubleshooting High in CLAT,
low in PLAT High
Security The security issues and concerns. DoS attack on the
binding table of
It dose not
Traffic hijacking could happen by
man -in-the-middle attack on
DHCPv6 which provisions the rules.
functions, no one translation mechanism can be used in all transition scenarios. This paper has provided the first
applicability index system, and highlighted the applicability of all IPv6 trans lation mechanisms to help the oper-
ators decide on the development scheme for th eir IPv 6 transition .
We can observe from the applicability analysis that 1) n et wo rk -side tran slation leads less overhead to ISP
than host-side translation, an d the mainstream of tra nslation techniques is n e two rk-s ide tr anslation ; 2) stateless
network -side translation requires voraciou s consumption of IPv4 addresses , while stateful translation requires
per-flow state maintenance, and both are not suitable for large scale networks; 3) Among the network-side
translation mechanisms, IVI is a feasible stateless tr ans lation mechanism, and NAT6 4 is a feasible stateful
translation mechanism; 4) 464XLAT is able to resolve the heterogeneous traversing problem of IPv4 mobile
host to IPv4 Internet, and MAP -T is more applicable to the heterogeneous traversing between IPv4 networ ks.
We are grateful for Jinggu o Ge fro m the Computer Network Information Center of Chin ese Academy of
W. Mi, X. D. Zhang
Sciences, who gave us kind encouragement and useful instruction s all through o ur wr iting.
Furthermore, this work is supported by the National High-tech R&D Program of China (863 Program) under
Grant No. 2011AA01A101; the National Key Technology Research and Development Program of the Ministry
of Science and Technology of China under Grant No.2012BAH01B00; the Nationa l Program on Key Basic Re-
search Project of China (973 Program) under Grant No. 2012CB315800; the Strategic Priority Research Pro-
gram of the Chinese Academy of Sciences under grant No. XDA06010201.
 AlJa’afreh, R., Mellor, J. and Awan, I. (2009) A Comparison between the Tunneling Process and Mapping Schemes
for IPv4/IPv6 Transition. International Conference on WAINA’09, IEEE Press, Bradford, 601-606.
 Law, Y.N., Lai, M.C., Tan, W.L. and Lau, W.C. (2008) Empirical Performance of IPv6 vs. IPv4 under a Dual-Stack
Environment. IEEE International Conference on ICC’08, IEEE Press, Beijing, 5924-5929.
 Aazam, M., Syed, A.M., Khan, I. and Alam, M. (2011) Evaluation of 6to4 and ISATAP on a Test LAN. IEEE Sympo-
sium on ISCI, IEEE Press, Kuala Lumpur, 46-50. http://dx.doi.org/10.1109/ISCI.2011.5958881
 Gilligan, R. and Nordmark, E. (1996) Transition Mechanisms for IPv6 Hosts and Routers. IETF RFC 1933.
 Guerin, R. and Hosanagar, K. (2010) Fo st e r ing IPv6 Migration through Network Quality Differentials. ACM SIGCOMM
Computer Communication Review, 40, 17-25. http://dx.doi.org/10.1145/1823844.1823847
 Wu, Y. and Zhou, X. (2011) Research on the IPv6 Performance Analysis Based on Dual-Protocol Stack and Tunnel
Transition. 6th International Conference on ICCSE, IEEE Press, Singapore, 1091-1093.
 Jayanthi, J.G. and Rabara, S.A. (2010) Transition and Mobility Management in the Integrated IPv4 and IPv6 Net-
work—A Systematic Review. International Conference On ICEIE, IEEE, Kyoto, 151-162.
 Wu, P., Cui, Y., Wu, J.P., Liu, J. and Metz, C. (2012) Transition from IPv4 to IPv6: A State-of-the-Art Survey. IEEE
Communications Surveys and Tutorials, 99, 1-18.
 Wu, P., Cui, Y., Xu, M., Wu, J., Li, X., Metz, C. and Wang, S. (2010) PET: Prefixing, Encapsulation and Translation
for IPv4-IPv6 Coexistence. GLOBECOM 2010, IEEE Press, Miami, 1-5.
 Miyata, H. and Endo, M. (2010) Design and Evaluation of IPv4/IPv6 Translator for IP Based Industrial Network Pro-
tocol. 8th IEEE International Conference on INDIN, IEEE Press, Osaka, 142-147.
 Govil, J., Kaur, N. and Kaur, H. (2008) An Examination of IPv4 and IPv6 Networks: Constraints and Various Transi-
tion Mechanisms. IEEE Southeastcon, IEEE Press, Huntsville, 178-185.
 Ge, J.G., Mi, W. and Wu, Y.L. (2014) The IPv6 Transition Mechanisms: Survey, Evaluation Criteria and Deployment
Considerations. Journal of Software, 4, 896-912.
 Nordmark, E. (2000) Stateless IP/ICMP Translation Algorithm. IETF RFC 2765.
 Li, X., Bao, C.X., Chen, M., Zhang, H. and Wu, J. (2011) The CERNET IVI Translation Design and Deployment for
the IPv4/IPv6 Coexistence and Transition. IETF RFC 6219.
 Tsirtsis, G. and Srisuresh, P. (2000) Network Address Translation—Protocol Translation (NAT-PT). IETF RFC 2766.
 Bagnulo, M., Matthews, P. and Van Beijnum, I. (2011) Stateful NAT64: Network Address and Protocol Translation
from IPv6 Clients to IPv4 Servers. IETF RFC 6146.
 Huang, B., Deng, H. and Savolainen (2012) Dual-Stack Hosts Using “Bump-in-the-Host” (BIH). IETF RFC 6535.
 Mawatari, M., Kawashima, M. and By r ne, C. (2013) 464XLAT: Combination of Stateful and Stateless Translation.
IETF RFC 6877.
 Bao, C., Li, X., Zhai, Y., et al. (2013) dIVI: Dual-Stateless IPv4/IPv6 Translation. IETF Draft (Work in Progress).
 Li, X., Bao, C., Dec, W., et al. (2013) Mapping of Address and Port Using Translation (MAP-T). IETF Draft (Work in