The Signpost

Technology report

To support or not to support IPv6, and why knowing when this report was last updated might be getting easier

Contribute  —  
Share this
By Jarry1250

IPv6 rolled out

As previewed last week, support for version six of the Internet Protocol (normally known by its initialism "IPv6") was enabled on Wikimedia wikis on June 6, hyped as World IPv6 Launch Day. IPv6 succeeds the widely-used IPv4 form that most people are familiar with, replacing the common IPv4 address (like 93.72.7.12) which can only provide 232 = 4,294,967,296 unique addresses with a longer 128-bit hexadecimal string (such as 2001:0:4137:9E76:247C:A71:833A:FA41).

The change, which is slowly being made by website providers around the world, will eventually allow for far more than 4.3 billion devices without introducing the potential for collateral damage occurring when an IPv4 address comes to represent many users (using NAT). By comparison, the Internet is projected to grow to 15 billion active devices by 2015; whereas this would have posed a problem under IPv4, IPv6 has been deemed sufficiently broad to offer the Internet almost unlimited room to grow.

While only a very small fraction of anonymous edits now come from IPv6 addresses, the June 6 deployment has caused significant disruption. Various scripts that are now being fed IPv6 addresses as input are either fully or partially broken due to the new format of the addresses. For example, Huggle was reported to choke on IPv6 address edits, and popups does not yet recognise IPv6 addresses as valid anonymous users. Various Toolserver scripts need updating as well, especially WHOIS and other IP address lookup tools regularly used by Wikimedians to counter disruption. Fixes to the German Wikipedia's vandal fighter community tool infrastructure, built and run by a small group of volunteer coders on behalf of the whole community, are expected to take weeks.

I get that this was an exciting step for the engineers who got it done, and I tip my hat to all of them for pulling it off; from that sense it's been a successful implementation [but] I also get that at least 30% of WMF users on hundreds of projects – that's roughly how many use one or more gadgets, scripts or tools that didn't work after this switch – have now had their "editing experience" negatively affected, and that almost all of it could have been avoided with a month or two of notice.

—English Wikipedian User:Risker. Responding, system administrator Ryan Lane asked whether that many tools had in fact been as badly affected as she had implied.

Even so, the disruption was considerably less than would have been experienced last year, when the Wikimedia Foundation had to drop out of World IPv6 Day because some parts of its database were not ready to accommodate IPv6 addresses. Indeed, this time around the issues seemed to have been successfully resolved by the World IPv6 Launch on June 6, if only just.

Despite the successful switch-on itself, the deployment has been far from uncontroversial: since June 6, there has been substantial criticism of how late in the day the Wikimedia Foundation seemed to resolve to take part in the launch event: right up until an announcement several days before, there had been numerous conflicting rumours about the WMF's participation, based on a few vague words by system administrators here and there. The lack of a Wikimedia Foundation listing at the World IPv6 Launch website further clouded the picture.

Unless extremely serious issues arise, it is planned that IPv6 will be enabled indefinitely. The new protocol poses a learning curve for administrators; at least three administrators on the English Wikipedia, for example confused IPv6 addresses with accounts on World IPv6 Launch day itself. It also poses a complication to CheckUser functionality. Fortunately, there is still time to learn, because IPv6 users present an extremely small minority (less than 0.7%) of editors on Wikipedia; the vast majority of IP and account blocks are still for IPv4 and will be for some time.

In brief

Signpost poll
Tech events
You can now give your opinion on next week's poll: Which of these best sums up your view about the move to Git and Gerrit?

Not all fixes may have gone live to WMF sites at the time of writing; some may not be scheduled to go live for many weeks.

+ Add a comment

Discuss this story

These comments are automatically transcluded from this article's talk page. To follow comments, add the page to your watchlist. If your comment has not appeared here, you can try purging the cache.
I'm confused. Why would we use ajax (or js whatsoever) to put lastmodified text on to a page. Its not like such text is dynamic, so caching shouldn't really be a concern [unless we wanted to change the timestamp magically before the page even reloads, which I somewhat doubt]. Not to mention there already is a timestamp in the lower left corner of every page... Bawolff (talk) 18:35, 12 June 2012 (UTC)[reply]
Oh wait, was it the clicktracking that caused the overload, that would make more sense. Bawolff (talk) 18:42, 12 June 2012 (UTC)[reply]
Yes, it was a simple logic error in clicktracking that measured far too many impressions. A patch has been submitted, and it won't be deployed until that's been merged in and thoroughly tested. Steven Walling (WMF) • talk 18:46, 12 June 2012 (UTC)[reply]

Could someone please modify the story to link the contribs of an IPv6 anonymous user? I think this would be really interesting to see and drive the point home. Dcoetzee 19:40, 12 June 2012 (UTC)[reply]

There is already a link, last time I checked.--Jasper Deng (talk) 01:10, 13 June 2012 (UTC)[reply]
I added one after Dcoetzee posted but forgot to respond here, just to clarify. - Jarry1250 [Deliberation needed] 09:48, 13 June 2012 (UTC)[reply]

Something's wrong with this week's pie chart summarizing last week's poll result. Light blue and yellow both show 22% but the description says 3 voters and 7 voters respectively. And red says 16% on the pie but got more response (9) than light blue or yellow. Please recheck the numbers. OhanaUnitedTalk page 18:18, 14 June 2012 (UTC)[reply]

Oh, good spot. The numbers were correct, but the pie chart not. We'll run the same chart next week with a published correction, I think. - Jarry1250 [Deliberation needed] 20:51, 14 June 2012 (UTC)[reply]



       

The Signpost · written by many · served by Sinepost V0.9 · 🄯 CC-BY-SA 4.0