Suggestion for Internet search engines: “Proposed IPv6 impact on search engine scoring algorithms”

At IETF76 in Maastricht I had a long discussion with Sander Steffann, IPv6 advocate and RIPE Address Policy WG co-chair. We talked about my idea, how could Google and all other search engines really trigger massive transition of internet content (and content providers) to IPv6.

We would like to invite influential end experienced decision makers, working at Internet search engines (specially *the search engine*) to read this proposal carefully – so if somebody knows well one of this guys, feel free to point them at following text:

Deployment and adoption of IPv6 is slow at this point in time. This can be a risk for future growth of the internet. One of the observed obstacles is that content providers are waiting for IPv6 viewers, and viewers are waiting for IPv6 content. We propose a slight change the search engine scoring algorithms to stimulate content providers to make their content accessible over IPv6.

Proposed change
We propose that search engines check whether a website is available over both IPv4 and IPv6. Having the same content available over both lower level protocols is the situation that will give the IPv4 to IPv6 transition the largest chance of succeeding. The way to determine if a website has ‘good’ IPv4 and IPv6 support is an implementation detail of the search engine.

Websites that are available over both protocols should then get some kind of bonus when compared to websites available over only IPv4 or IPv6. One possible idea is to use this as a tiebreaker when two pages get the same score based on the original scoring algorithm. Another possibility is to give the website a 5% bonus in the scoring algorithm. This choice is an implementation detail of the search engine.

The search engine operator should then make it publicly known that IPv6 support will have a positive impact on the search engine scoring algorithm.

This will stimulate website owners to make their websites available over IPv6, which benefits the whole internet community. For cases where the website owner makes use of services from a separate website hoster this hoster will also be stimulated to support IPv6. It will also send a signal that the search engine operator sees IPv6 support as being important for the future of the internet. Improving the future internet is also in the best interest of the search engine operator itself, as their business is based on the content available on the internet.

This proposal changes the scoring algorithm of the search engine, which is a very important part of the quality that the search engine provides. Using IPv6 support only as a tiebreaker or as a small component in this algorithm minimizes the impact. Another con can be that the public will see the search engine operator as pushing a technology instead of focussing on returning the search results that are ‘best’ for the end user. A counter argument to this can be that making content available over IPv6 is in the long term in the users best interest.

Thanx to Sander for taking my thoughts into text, I think it’s clear enough as is. With this move, everybody is a winner, nobody looses anything – “Search engine” is considered user friendly, taking care of their future experience and content providers would start rushing to deploy IPv6 to be just a step in front of their competitors. Done deal ­čÖé

Comments and debate warmly welcome!

Jan Zorz,

Vaš IP naslov (ali ste na IPv6 ?):


  1. Ragnar6
    August 23rd, 2010 | 11:01

    Podpiram to idejo.

  2. Phil
    August 23rd, 2010 | 13:13

    I forsee a problem as this inherently requires dual stacked www (et al) records and dual stacking is famously ropy where end users have broken or partial v6 implementation/visibility as is, sadly, very often the case.

  3. August 23rd, 2010 | 16:24

    @Phil: As Lorenzo figured out, 0,2% (or something like that) of users may be impacted. Wouldn’t care…

  4. August 25th, 2010 | 09:25


    Great idea and good work putting it together.

    It seems to me to be a reasonable proposal given that search engines already take into account factors other than the content of the web-site.

    Of course I am biased as our web-sites are all IPv6 enabled!


  5. August 25th, 2010 | 09:35

    David: Yes, transport factors are already taken into account…

  6. Gert Doering
    August 25th, 2010 | 17:24


    great idea. Go for it!

    (It is *important* to have dual-stack content *now* – so that the remaining issues can get fixed in time. Waiting until the big boom is not going to make the transition any smoother).


  7. August 26th, 2010 | 10:33

    […] got massive feedback on proposed change, mostly very positive and some of brave fellows also left pubblic comment below the article […]

  8. August 27th, 2010 | 10:15

    […] to the fore. As a result, Jan Zorz and others have put together a proposal which can be seen at, ┬á…. This proposal recommends the inclusion of IPv6 in the factors used by search […]

  9. August 27th, 2010 | 23:00

    My thoughts on possible negative impact of the proposed change:

    “Search engine is serving URLs in SERP regardless if they are A or A/AAAA recorded in DNS, so rewarding of A/AAAA record availability can’t currently impact on users search experience, but just reward content provider serving content on IPv4 and IPv6, making it possibly future proof.”

    Hope that this thought goes into internal debates inside search engines ­čÖé

  10. ipv1
    September 1st, 2010 | 22:53

    I first was a bit skeptic to use ranking to stimulate IPv6 growth. With normal normal native transport user experience to a particular website should be the same for IPv4 and IPv6. But when you think of the recent IPv6 trail of T-Mobile where they use NAT64 this changed for me. Users in this trial will get better performance when the web site that they are trying to reach is dual-stack. And when I was reading about stuff what Google is using to influence page ranking, performance is something they care about

    I like the Google whitelist but search engines should announce AAAA records (without whitelists) before doing this. They can’t say that they don’t want 0.1% users with problems but we expect that you take it for granted. No flam..

    Is this the businesses cause for IPv6?

  11. September 22nd, 2010 | 16:36

    […] O pessoal do go6, da Eslov├¬nia, tem uma proposta muito interessante para melhorar a ado├ž├úo do IPv6 na Web. Bastaria que o Google e outros buscadores passassem a utilizar o IPv6 como um dos crit├ęrios para calcular o ranking das p├íginas. Ou seja, se a s├ştio Web responder corretamente em pilha dupla, ele aparece um pouco melhor colocado na resposta do buscador. Me parece algo que valeria a pena apoiar:… […]

Leave a reply