Microsoft offers UK cloud customers private pipes

Microsoft is offering private connections to its UK data centres as a means of wooing government customers to its sometimes less-than- resilient cloud. The private connection means faster pipes, less latency and therefore less failover or loss of service. Also, supposedly, greater protection. Until now, Microsoft relied upon …

[Anonymous Coward]

Even better!

Routing via GCHQ is compteletly transparent.

[adnim]

A tad presumptuous?

"The private connection means faster pipes, less latency and therefore less failover or loss of service. Also, supposedly, greater protection."

After all... This is not just a cloud, this is a M&S cloud

[Anonymous Coward]

Re: A tad presumptuous?

"The private connection means faster pipes, less latency and therefore less failover or loss of service"

"Nicer ambulances, faster response times and better-looking drivers mean they’re not just ‘the’ emergency services – they’re ‘your’ emergency services. So, remember the new number ..."

[Dazed and Confused]

less latency and therefore less failover

OK, so if they know that little about networking then I can think of a really good reason for not going to talk to them.

Perhaps this explains their recent hiccough with DHCP.

[Anonymous Coward]

Re: less latency and therefore less failover

Not strictly true. If you are dependent on a latency sensitive application such as video or voice e.g. Skype, if the latency goes to high, you may well switch over to a different system / site / circuit.

[Tim Seventh]

Re: less latency and therefore less failover

You missed the article, didn't you?

DON'T MISS: Microsoft quietly emits patch to undo its earlier patch that broke Windows 10 networking

http://www.theregister.co.uk/2016/12/14/microsoft_windows_10_dhcp/

[Dazed and Confused]

Re: less latency and therefore less failover

The difference in latency of using a direct link and leased fibre via an ISP is likely to be tiny. A leased fibre is giving you what? a 2ms latency via an ISP, at least that's what I used to see on a line from BT to a data centre in the middle of nowhere going to some other Internet sites over 10 years ago. Skype runs happily over comparatively large latencies or we wouldn't all use it. I'm not an expert on the human perception in communication, but I'd be surprised if you can spot a 100ms delay. delays of over 200ms are human recognisable but don't seem to affect how easily we can talk to people in Oz, delays of a whole second mess you up and you have to get used to not interrupting each other.

So I stand by my comment that it's generally wrong to associate latency to availability.

And no I didn't miss the story about MS having patched their patch, but if you can't get on the network you can't load the patch for the patch.

[allthecoolshortnamesweretaken]

So, back to dedicated lines for the important stuff?

Let's party like it's 1999 1966!

[Deltics]

We already were... Having gone back to paying to use somebody else's computer in a data center.

The only question is how long it will take for this cycle to complete and for some bean counter somewhere to announce the revelation that there are huge cost and efficiency savings (and reliability and control gains) to be had by de-centralising.

Rinse and repeat.

[Doogie Howser MD]

Outage Stats

It's probably worth making the distinction between O365 and Azure - O365 seems to be on it's arse more often than not, but to the best of my knowledge, Azure has been pretty sound.

I know they're all shoved under the marketing umbrella of "Microsoft Cloud" but they're very different beasts, really.

[Just Enough]

Re: Outage Stats

You think that they're not physically based in the same locations using the same networks?