02-04-2013 09:12 PM - edited 03-01-2019 05:38 PM
Hello
Please could anyone help me with IPv6 transision issues , I'm doing a research and its about to find IPv6 transition techniqeues issues , so for example tunneling or translating could anyone give me any suggestions where i can find anything related to that problems , thank you so much
Solved! Go to Solution.
02-06-2013 10:07 AM
If you want to take this discussion off the cisco support forum, my personal e-mail address is jeleinweber@gmail.com.
I'd suggest looking at NAT64+DNS64 or its predecessor IVI if you want to investigate transition mechanisms academically.
Since it takes about a decade to create something, implement it, standardize it, get vendors to ship it, train staff to configure it, and get it widely deployed, no brand new transitions technologies are likely to be involved in the v6 migration. The recommended and observed behavior is that all v6 access is going native; all clients going foward are going to be either dual-stack or v6-only; transition tunneling technologies for clients like Teredo & 6to4 are being both abandoned and deprecated. E.g. nearly all v6 traffic into google is now native v6.
For ISP's which haven't completed their dual-stack rollout, they are papering over their remaining v4-only islands using 6rd at the edge and 6PE on the backbone, and those are fairly simple and fairly mature and temporary, making them uninteresting to study.
So the remaining big transition issue is how the growing number of v6-only clients will access the legacy v4 internet for the next 4-6 years. Currently the number of such clients is small; 3 years from now the number may be large. Think 2-3 billion v6-only devices in 2015. The long tail of v4-only web servers will probably still be 70% of the DNS names at that point, even if they only get 5% of the backbone traffic.
-- Jim Leinweber
02-05-2013 08:14 AM
IPv6 transition issues vary a lot depending on the specifics of one's situation. If you describe yours in more detail you can get more specific advice. E.g. are you researching on behalf of end users, organizations, vendors, or ISP's? Individuals should be buying only dual-stack modems and wifi gear, e.g. DOCSIS 3 cable modems. Organizations should be demanding dual-stack native support from their ISP. Vendors should be certifying their v6 stacks with the US NIST or DOD profiles. ISP's should be using 6rd and 6pe tunnels to carry end-to-end v6 over the v4-only regions of their network while they rush their native universal v6 support to completion.
Since IANA, APNIC, and RIPE have all run out of IPv4 allocations, legacy v4 internet use for a large part of the planet is headed toward Carrier Grade NAT44, with all of the usual NAT44 problems (see rfc4966) plus the likelihood of double-NAT problems plus all the fun of sharing your NAT problems with lots of neighbors, some of them surely botnet zombies. Just the legally required logging is a nightmare in regions like the USA under CALEA. Meanwhile, at least in the USA, all of the up and coming LTE4 cellphone data networks are v6-only, meaning you have to go through a carrier NAT64+DNS64 gateway to get to the legacy v4 internet. The cellphone companies won't care about lousy v4 performance with the bottom 95% of the Alexa top million sites; the high traffic destinations like facebook and youtube and google are already native on v6.
There are way too many tunneling technologies, most of them deprecated, like 6over4. In general, the "automatic" ones like Teredo (2001:0::/32) and 6to4 (2002::/16) which depend on remote 3rd party relays have nasty latency problems and horrible reliability problems, and should be avoided. The best v6 is native. Next best is 6rd, where there is an ISP protocol 41 (slap a v4 header on the front of your v6 packet) tunnel between the ISP modem and an in-AS ISP relay, converting to full native v6 at the ISP border. For R&D purposes a 6in4 tunnel, still protocol 41, either with your ISP or with a tunnel broker like Hurricane Electric, Sixxs, or gogo6 is your best bet.
Geof Huston, chief technologist for APNIC has a lot of v6-related stuff at potaroo.net. Hurrican electric has a useful site at ipv6.he.net. Both ARIN and RIPE have IPv6 pages with links to some resources. Ars Technica has run some good IPv6 articles over the years; the best on on the ugliness of transition technologies is probably "There is no plan B":
I've done a number of IPv6 tutorials around the state of Wisconsin; slide and lab exercise materials for the most recent one are at:
https://mywebspace.wisc.edu/jeleinwe/web/ipv6/det/
-- Jim Leinweber, WI state lab of Hygiene / University of Wisconsin - Madison
02-06-2013 01:40 AM
Dear James
I'm doing last year project of my degree , so the question I'm trying to answer is , how to improve IPv6 transition mechanisms , so i can take any particular mechanism , but I'm just confussed which mechanism i should investigate and how to choose it ,
thank you for your comment I'm glad , and i will be more than happy if we can have contacts , in order to discuss with you if you can help me further more.
02-05-2013 03:11 PM
Hello Ibrahim,
IANA Pool (www.iana.com) has been depleted on February 3rd 2011, and the last blocks assigned to the regional Internet registries is decreasing at a significant rate, so their depletion is rapidly approaching. In fact the APNIC one was already exhausted on April 2011.
The Current Protocol (Internet Protocol Version 4 or IPv4) supports approximately 4 billion addresses, and due to the huge success of the internet, it’s anticipated that these will be depleted soon.
It’s important to highlight the fact that, although NAT has so far allowed the internet to continue growing, this technology implies the loss of end-to-end connectivity and therefore, hinders the deployment end-to-end (client-to-client) application and services, making the development of those services and applications more complex.
The new IPv6 Protocol Supports 340 Trillion trillion (six trillion) addresses, which makes the number of IPv4 addresses appear significant. Along with this Larger address apace, IPv6 offers a varity of advantages for network administration in terms of stability, flexibility, simplicity and availability.
Gradually, IPv6 networks are being implemented and, during this transition, IPv6 and IPv4 will coexist for many years to come. Although most of the Technical work having to do with the protocol has already been completed. IPv6 is required to be deployed in Internet Service Provider Networks.
I would therfore dont recommend translation of v6, although NAT64 and NAT-PT were introduced to resolve such requirements, but it would always be better to have a native v6 implementation! and the Huge Number of available v6 addresses makes it easy to achieve!
On the other hand, Tunneling v6 over v4 is a possible approach to establish end to end v6 connectivity.
Regards,
Mohamed
02-06-2013 01:40 AM
Dear Mohamed
I'm doing last year project of my degree , so the question I'm trying to answer is , how to improve IPv6 transition mechanisms , so i can take any particular mechanism , but I'm just confussed which mechanism i should investigate and how to choose it ,
thank you for your comment I'm glad , and i will be more than happy if we can have contacts , in order to discuss with you if you can help me further more.
02-06-2013 10:07 AM
If you want to take this discussion off the cisco support forum, my personal e-mail address is jeleinweber@gmail.com.
I'd suggest looking at NAT64+DNS64 or its predecessor IVI if you want to investigate transition mechanisms academically.
Since it takes about a decade to create something, implement it, standardize it, get vendors to ship it, train staff to configure it, and get it widely deployed, no brand new transitions technologies are likely to be involved in the v6 migration. The recommended and observed behavior is that all v6 access is going native; all clients going foward are going to be either dual-stack or v6-only; transition tunneling technologies for clients like Teredo & 6to4 are being both abandoned and deprecated. E.g. nearly all v6 traffic into google is now native v6.
For ISP's which haven't completed their dual-stack rollout, they are papering over their remaining v4-only islands using 6rd at the edge and 6PE on the backbone, and those are fairly simple and fairly mature and temporary, making them uninteresting to study.
So the remaining big transition issue is how the growing number of v6-only clients will access the legacy v4 internet for the next 4-6 years. Currently the number of such clients is small; 3 years from now the number may be large. Think 2-3 billion v6-only devices in 2015. The long tail of v4-only web servers will probably still be 70% of the DNS names at that point, even if they only get 5% of the backbone traffic.
-- Jim Leinweber
02-06-2013 03:35 PM
Dear James
sir your discussion is very useful and it helps me a lot , so the question i have for you is , for example if i choose to investigate NAT64 the reason will be because this mechanism we use to connect the IPv6 clients and the IPv4 servers, because v4 still some servers use it . give me little brief about the good reason why i choose NAT64
thank you again sir , and i will contact you through your email if i sometimes need your discussion urgently
02-08-2013 08:22 AM
Dear James
I've tried to reach you through your email and you didn't reply , i hope you doing fine , I'm glad to tell you that your discusion helps me a lot , i decided to investigate NAT64/DNS64 so please any recomendation how could you help me at this point
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide