Monday 23 June 2014

"Superfast Broadband"

Business Wales have sent around an email making local businesses aware that "superfast broadband" is being rolled out across Wales and pointing out various benefits that businesses should look into.  They seem to be talking about FTTC and FTTP connections (they don't mention how available FTTP is - I think "not very" may well be accurate).

This is all quite sensible to make businesses aware of what's going on since the less technical businesses may well not be aware of this stuff.  However, the material they've sent out looks a bit questionable to me.  For starters, they've linked to a savings calculator which works out how much you could save by shifting some of your internal equipment out to "the cloud".  One of the provisos they give is "If your organisation hosts more than 1 server it will use it purely for file, print or authentication purposes" but they don't mention anything about what services you're using if you only have a single server - I'm pretty sure their pricing doesn't cover a lot of the services we run on our servers - i.e. one of our servers runs an Asterisk phone exchange, Subversion/Trac revision control, support ticketting system, a name server, email service, etc.  We also have several other servers doing different things, but in principle we could consolidate a lot more stuff onto a single server if we wanted to.

But that proviso also listed some slightly insane stuff - why would you want a print server to be in "the cloud"?  Print jobs can be *huge*, so sending them up over the internet to a print server, only to be pulled down again to the printer seems nuts to me.  Also, your printer is going to need to be connected to the network in order to talk to the print server, so is there actually much benefit in having the workstations talk to a separate print server instead of just sending jobs directly to the printer?

The other thing that caught my eye was that they are saying that servers hosted in the cloud are free.  They are suitably vague about where they're these free services come from - Google Apps, for example, is £33/user/year - certainly not free.

They do talk about the SLA provided by the cloud services, quoting figures like 99.9% uptime, but what about the reliability of the internet connection itself.  This certainly shouldn't stop people from considering cloud services, but you do need to consider that if your business is wholly reliant on a fast internet connection for everything then you're going to have serious problems if that connection is down for a long time.  BT have been known to take several days to repair faults, and whilst you can pay for an SLA that "guarantees" a fast fix, the fact remains that BT frequently miss the guaranteed fix times and the compensation is a pittance (they quote £25 compensation if they don't meet the SLA).  A small office can get by with a 3G connection for a few days if they are just using the internet for some web surfing and email; but this isn't going to work if you're expecting to shift move many gigabytes of files and print jobs between your network and a cloud service.

There's also some blurb about how to choose an ISP, but it seems to muddle lots of bits and pieces together here.  They talk about the ISP hosting your domain, email, website but don't (in my opinion) make it clear that these things might already by hosted elsewhere.  There is no discussion about why hosting these things with the ISP may, in fact, not be a great idea (e.g. tieing your web hosting into your internet service may make changing either one of them in the future quite tricky).

Finally, there's absolutely no mention about IPv6 connectivity.  This is going to become a big deal - the only part of the world that has IPv4 addresses left is Africa, and they are forecast to run out in around 5 years; for everyone else, there are no spare IPv4 addresses - any time an address is needed it has to be repurposed from another service.  There are a few ways of dealing with this problem, but fundamentally everyone needs to migrate onto IPv6.

Pretty much every ISP is going to need to implement either CGNAT or NAT64 as a stop-gap measure to ensure IPv4-only services remain accessible.  However, these technologies are bodges, each with their own list of problems and the only real solution is going to be to move everything to IPv6.

Some ISPs, such as A&A and EntaNet have been providing IPv6 connectivity for years and if you're on one of these and have a properly configured router you don't need to worry about it at all.

Some other ISPs have said they won't be rolling out IPv6 any time soon - for example, PlusNet have said they are going to use CGNAT instead.  As mentioned above, CGNAT isn't really a solution, it's a stop-gap measure with a whole host of problems, so prolonging the use of the stop-gap by not rolling out IPv6 seems extremely questionable.

Most ISPs have remained completely quiet on the issue, and anyone using these ISPs should certainly be concerned since they have provided no indication about how well they have been planning for the future.  To be clear: all ISPs have had almost 20 years to plan and roll out IPv6 infrastructure, so there isn't really any excuse for this not to be well under way.

Friday 13 June 2014

IPv4 running out

Well, it was bound to happen sooner or later - Microsoft say they have
run out of US IPv4 addresses for their Azure service and are now starting to assign non-US addresses to US services.

This follows LACNIC (the regional registry responsible for Latin America and the Caribbean) announcing that their Gradual Exhaustion and New Entrants policies had come into effect on Monday as they hit their last /10 (around 4 million IPv4 addresses).  This essentially means that the remaining addresses are going to be allocated on an extremely strict policy, which to all intents and purposes means there are no more addresses to be had from LACNIC.

APNIC (Asia Pacific) and RIPE (Europe) have been running under their respective end-game policies for some time now, and ARIN (North America) are also operating under their "phase 4" policy, although this policy seems very vague compared to those of the other RIRs.  So the only region where IPv4 addresses are still being handed out is AFRINIC (Africa).

AFRINIC have a relatively large number of addresses left and don't really seem to burn through them very fast, probably owing to relatively slow technological development within the region - they have about 3.14 /8s (about 53 million addresses) and are projected to run out some time around the middle of 2019 with their current rate of consumption.  It will be interesting to see if the AFRINIC consumption now increases as global businesses, such as Microsoft, start requesting assignments from AFRINIC to be used in other regions.  Indeed, I wonder if the history of the applicant will count against them - will AFRINIC refuse to hand addresses to Microsoft, for example, because they now have a history of using large numbers of non-US addresses on US services?

I note that The Register has made a few silly comments in their article on this latest news that probably need some clarification:

Firstly they ask "why Azure is relying on Ipv4".  Microsoft's blog post didn't mention anything about the internals of Azure - it didn't indicate that the actual infrastructure was affected at all.  Indeed, the internals are probably running on IPv6, or private IPv4 addresses, so wouldn't be affected by the IPv4 address shortage.  However, The Register has failed to figure out that the people hosting services on Azure want the general public to be able to access them, and with the majority of ISPs still dragging their feet to roll out IPv6 to their customers there's no alternative but to use IPv4 on the server end too.

The other question they pose is "how Redmond let itself run out of IPv4 addresses" when they were "buying IPv4".  Now to be clear: IPv4 addresses have never been a purchasable commodity.  Addresses were assigned by the regional registrars according to strict criteria - at no point could Microsoft have gone to a registrar and said "can we have enough addresses to last us 10 years please" - the registrar would've told them to get lost (RIPE's policy was for LIRs to only get 6 months' worth at a time, for example).  Now, MS did, in fact, buy a bunch of IP addresses from the bankrupt Nortel in 2011, and there was some surprise that the registry didn't just demand that they be returned since strictly speaking this kind of sale doesn't seem to be allowed - the terms of most registries is that you return addresses to the registry when you stop using them.  Arguably a company that is selling off thousands of addresses is no longer using them, so they should've probably been returned rather than being sold.  In any case, in the general case, IPv4 addresses have never been for sale, so its hard to fault Microsoft for not buying up the world's supply years ago.

What is clear, however, is that everyone, especially ISPs, seriously needs to pull their finger out and roll out IPv6 soon.  Its long past the point where we can expect a slow trouble-free migration, we're well into a period where there's going to be some major disruption as people who have been dragging their feet suddenly spot the sky falling in on them.

Monday 9 June 2014

Npower, yet again

So after Npower's previous screwup, they've done it again.

To recap, In January 2013 I switched my gas and electricity to Npower.  They set up the direct debit and everything, but sent me no paperwork at all until seven months later when I got my first bill... and it was enormous.  Coz as it turned out they had put me on completely the wrong tariff and hadn't actually bothered to charge the direct debit (I hadn't noticed).  It was merry hell to sort it out and took a considerable amount of my time, with them trying to the deflect blame back to me at pretty much every step (e.g. yes they had forgotten to charge the direct debit, but apparently that was my fault, because I'm apparently responsible for checking up on them all the time to ensure they've done their job.  I do wonder what the advantage of direct debit is supposed to be over just paying manually every month if they think I have to check it every month anyway).

Anyway, it eventually got sorted out and I paid off the hundreds of pounds of debt that had amassed on my account.

In January this year, I got a letter from them saying I was overpaying and that they were reducing my DD payments by about half.  All well and good.  However, I now have a letter telling me I am hundreds of pounds in debt to them (again) and that I must pay this within 9 days of the letter date (actually, I must pay it by 26th May, but of course I didn't get the letter on time because I was away from home).  It turns out that rather than reducing the DD payments, they just plain stopped them altogether (again).

So clearly its their fault that I've ended up with a load of debt against my account again (although I'm sure they will insist that its my fault), but even if it wasn't their fault, how is it right that they first thing they've done to let me know is give me 9 days notice to pay up?  Surely if the account is set to be paid by DD but isn't actually being paid by DD then they should tell the customer immediately rather than waiting until the account is a few hundred quid in debt?

Needless to say, I've now switched energy supplier and am waiting for Npower to call me back to reach a resolution for this...  Never before have I dealt with such an utterly incompetent company.

(Edit: Having just got off the phone with them, they have applied the £88 direct debit discount that they "forgot" to deduct from my account (despite me querying it in January already) and have let me pay off the balance so I can switch supplier.  I await the revised bill that they are sending through so I can check it's actually accurate this time.  I do wonder how much money they make off "forgetting" things like DD discounts and hoping that people don't notice.  Maybe I'm just cynical, but I feel like there isn't much incentive for them to fix their systems when the billing errors are making them money.)