[Gllug] Geographcial server failover

[Richard Cottrill]

> There's at least one more critical system than that that springs to
> mind: telco call-routing (and to a lesser extent usage-tracking) and
> exchange systems. If they fail, at the best the telco starts
> haemorrhaging money at titanic speeds; at worst, 999 calls fail, and
> people die.

I find myself in the novel position of not just disagreeing, but even
disbelieveing one of Nix's emails. It had to happen eventually.

<war stories>
Ah telcos, bless em. I have recent experience of a telco I worked for who
had regular failures of their billing system, constant in fact. It did go
right up the the CEO after a few months, and was eventually fixed by some
back room hacking by a friend of mine, much to the inconvenience of most
of the senior management who had just submitted a project proposal for
megabucks, over months, for a less effective fix.

Why is billing an issue like call-routing? If the billing system doesn't
think you credit worthy, your call won't even make it into the ether. 999
is exempt though, but I think the company in question has had issues with
that too.
<rants>
I would be shocked if BT *hadn't* had serious interruptions to service,
ever; my money's on recently. People talk about "telco-like reliability",
but in my experience, that's only words that don't translate into actions
within the telcos.

The IBMs, Accentures, and the rest (including my current employer) have a
lot to answer for in gutting the "top end" of IT of technical skills, and
replacing it with crap and spin. There are whole swathes of the industry
who don't understand what (or why) best (or even acceptable) practice is.
</rants>
</war stories>

Richard

-- 
Gllug mailing list  -  Gllug at gllug.org.uk
http://lists.gllug.org.uk/mailman/listinfo/gllug