In these large colo datacentres for cloud players, who looks after the hardware such as remote hands? Do MS, Amazon and (to a lesser extent) Google have one or two of their own chaps on standby for their respective suites, or do these suites literally never have anyone technical from said cloud owner ever see their kit?
The dark Satantic mills of the internet belched out more cash than they ever have for chunky data centre landlord Equinix in its calendar Q1. The company reported an 11 per cent year-on-year rise in turnover to $1.363bn for the three months ended 31 March: 28 per cent of this was generated by sales of rack space to cloud and …
Friday 3rd May 2019 10:22 GMT Mr Fix.
From what I've heard at conferences the biggest player have built their systems in such redundant way that they dont care if things break. A customer who had seen the insides of an azure DC said he was impressed by all the red lights everywhere (Microsoft spokeswoman on stage seemed quite uncomfortable with that statement). Then once in a while they sweep in an decomission whole racks at a time when its reached a critical amount of titsup.
Friday 3rd May 2019 12:08 GMT Anonymous Coward
I don't know all the details but I do know that the DC's are installed/upgraded using pre-built racks that are prepared offsite.
Reliability may be such that failures are dealt with in a similar way because the build out of a west London data center by everyone's favorite desktop provider consists of trucks delivering racks there most days.
Friday 3rd May 2019 15:22 GMT Anonymous Coward
I've worked for both AWS and MS. We always hire our own staff to look after the servers etc.
Depending on the site if it's just Colo space landlord looks after power etc and we do the rest inside the cage/Colo space
If it's a leased site (only customer) then we manage everything from power etc