So, there were a couple of large blocks allocated by RIPE yesterday – Etisalat (Saudi Arabia) had a /14, and an ISP in Germany had a /15. That’s another day of allocating 0.03 /8s in a day – we have days to go now if this continues, not weeks. The Huston-o-Meter is predicting 21 September as of today, but I’m not sure it’ll be that long if the allocation rate continues at the same pace.
Category Archives: Computers
The Daily IPv4 Report
A short report today, but amongst other smaller allocations yesterday, 256K addresses were allocated to Tele2 Sverige AB in Sweden, which brings us down another 0.03 /8s or so. If the allocation rates continue at that pace we might well be out by the beginning of next week …
This week in the IPv4 world
It’s been an interesting week in the world of IPv4. Still plenty of people trying to tell you that there’s no problem and no-one’s running out, but a couple of significant things have happened this week which are worth noting.
Firstly, RIPE NCC, the regional Internet registry for Europe made an official announcement last Tuesday (4 September) that they were down to their last /10 above the final /8, which is 0.25 of a /8 or 4,194,304 addresses. When this happens, RIPE move to ‘Phase 1’, which means that their allocation people work in pairs and scrutinise things much more closely. In addition, things are processed very strictly in order and they will only communicate with you when your ticket is in the front of the queue. RIPE have also made their official graph update daily now, which you can find at http://www.ripe.net/internet-coordination/ipv4-exhaustion/ipv4-available-pool-graph
The second thing to happen this week has come from a rather unusual place – Iran. Two different ISPs (who may or may not be government owned/controlled, I don’t know), asked for two blocks of addresses, one a /12 and another a /13 – a total of 1,572,864 in just one week. The sensation caused the “Huston-o-Meter” (at http://ipv4.potaroo.net) to jump from 4 October to 27 September and now stands, at the time of writing, at 22 September. If any more large allocations occur between now and then, we could be looking at a run-out date of less than 2 weeks time!
Good time to keep your head in the sand, isn’t it?
Status
Several 1980s home computers have celebrated their 30th or 35th anniversaries over the last week
Commodore 64 is 30 – http://www.bbc.co.uk/news/technology-19055707
Dragon 32 is 30 – http://www.reghardware.com/2012/08/01/the_dragon_32_is_30_years_old/
Tandy/Radio Shack TRS-80 is 35 – http://gizmodo.com/5931793/happy-35th-birthday-to-radio-shacks-trs+80-personal-computer-a-leader-in-its-day
Status
Here’s something not to buy in October
http://www.theinquirer.net/inquirer/news/2190251/microsoft-release-windows-october
One month to go?
So as I write, today is 29 June 2012. As of yesterday, 29 July 2012 was Geoff Huston’s prediction for RIPE to get down to the final /8 (although I notice it’s now at 28 July), at which point the supply of IPv4 addresses to Europe and the surrounding region effectively dry up, along with the APNIC region which “exhausted” last year.
The “head-in-sand” mentality still appears to be quite prevalent still, and I’m not sure that’s going to change until the day actually happens. Here’s hoping it’ll come soon …
World IPv6 Launch
So today, Wednesday 6 June 2012, one year on from the previous World IPv6 Day which was held on Wednesday 8 June 2011, is officially World IPv6 Launch Day. Major web sites such as Google, Yahoo, Facebook and others including Akamai, the content distribution network, have permanently switched on their IPv6 support today and, unlike last year, aren’t turning it off this time. I’ve no idea whether this will have any effect in driving IPv6 adoption, but we can only hope I suppose.
What was more worrying, but probably not surprising, was the large number of inaccurate press reports on the subject. The BBC News web site seemed to be telling us that IPv6 had only just been invented and switched on, when in fact it’s been around since 1997 in one form or another. Then there was the usual ‘who needs IPv6, just ignore it’ articles and lots more besides. It really doesn’t help when the press get it wrong like this!
My biggest disappointment of today was that, although Google had trumpeted how IPv6 compliant it was now, I thought I’d go for the obvious test and send an e-mail from Gmail to my own email address, which sits on an IPv6-enabled server. And guess what – although you can access Gmail’s web interface over v6, you can’t actually send the mail using v6 – looking at the headers, I was disappointed to see that the mail was delivered to my server over IPv4. That said, Google Talk (XMPP) is available over IPv6, so that’s good at least.
So, happy World IPv6 Launch day, and the next milestone is going to be RIPE running out of addresses, if the predictions are correct, in about two months’ time.
XMPP “fun”
Had lots of “fun” tonight with my XMPP-compatible instant messaging server, ejabberd. For some reason since I moved it between servers it has been misbehaving. This has been a “fun” one to track down, since it wasn’t obvious. But tonight I have found the answer. Erlang (the language that ejabberd is written in) is a little obscure at times, but it turns out that if you turn the logging up enough, then you see strange errors in the logs. After a bit of Googling, the answer turns out to be a bug in Debian. The ejabberd package depends on the erlang-xmerl package being installed, but Debian for some reason forgot to make that a dependency of ejabberd as described in Debian bug 670307. Hopefully they’ll fix that soon, and until they do, the fix is just to apt-get or aptitude install erlang-xmerl so that the package gets installed.
Why I’m blocking mails from Yahoo!
If you are sending e-mail to me from a @yahoo.com address or from a service provider that uses Yahoo! as its wholesale provider such as BT Retail Broadband customers (@btinternet.com), you are unlikely to be able to send to me any more. Why is this? Because Yahoo! have decided to block my very recently allocated IPv4 address range so that I cannot send mail to any Yahoo! customer or any of their wholesale customers – presumably because they think I’m a notorious Russian spammer or something (since the /16 that this block of addresses is in seems to contain blocks of IPs from places such as Russia and Kazakhstan). Which might be tempting to think, apart from the fact that my IPv4 addresses were blocked before I could even send them a single mail. Despite having proper SPF records and DKIM signing all outgoing mail, attempting to send mail to their servers results in a message thus:
421 4.7.1 [TS03] All messages from 178.238.157.69 will be permanently deferred; Retrying will NOT succeed. See http://postmaster.yahoo.com/errors/421-ts03.html
Yahoo will not explain why my IPv4 addresses are blocked, they sent a canned reply stating that they will do nothing about it, and that their spam blocking methods are so secret that they can’t (read: won’t) tell me why, with apparently no right of appeal. For obvious reasons, I consider this to be totally unacceptable behaviour on their part.
As far as I can tell, Yahoo! are the only major ISP who are blocking my address range, so using Hotmail, Gmail, etc. should work.
So, Yahoo, if you’re reading this – if you do not unblock my IPv4 address range on your servers (178.238.157.64/26), then your customers will not be able to send e-mail to me until you unblock them.
If you are a customer of an ISP directly affected by this, you may wish to take this matter up with them if you wish to carry on receiving e-mails from me.
Renumbering
Renumbering all your hosts is always “fun”. Today I’ve obtained a new range of addresses from my ISP, Andrews & Arnold, a shiny new /26 block of IPv4 (and in time before RIPE runs out, too.)
So for those of you that care about these things, you’ll notice I have a different IPv4 address range now (the IPv6 addresses are unchanged). Hopefully this won’t make a scrap of difference to anyone, but if you do experience problems, let me know.