[AusNOG] census issues tonight

[Robert Hudson]

On 10 August 2016 at 13:08, Roland Dobbins <rdobbins at arbor.net> wrote:

> On 10 Aug 2016, at 7:19, Mark Andrews wrote:
>
>  The public does know the difference between a denial of service attack /
>> overloaded server and a data breach.
>>
>
> Actually, most of them don't.  Nor do most journalists.
>

Nor do politicians such as Michael McCormack.  Or public servants such as
Duncan Young.

But apparently we should trust their claims that the servers would cope
with the load, and are totally secure, because they've never been hacked.

A colleague asked me to run a little exercise in critical thinking.  Here's
the output.  Makes you wonder what calculations the ABS were using when
they did their "planning" (I use the term loosely):

OK.  24m Australians.  2.1 people on average per household.  so roughtly
11.5m submissions.  Let's say 1.5m households do it either not on the night
or on paper, leaving us with a nice round 10m submissions to be done.  Some
assumptions - many of us finish work at 5pm, head home, have dinner, then
at around 7pm are ready to do other stuff (check TV scheduling to justify
this.  So, in, let's say 2 hours, we expected 10m submissions, or 5m an
hour.  You'd expect that to be slanted to the first hour, so let's say 6m
in the first hour, 4m in the second.
That reasonable?
Let's blow away my assumptions and say only 5m submissions were needed last
night.  That still leaves you with ~3m in the first hour, 2m in the
second...
And they planned for 1m an hour (evenly spread out I bet), and tested for
1.5 maximum.
Yep, planning fail.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ausnog.net/pipermail/ausnog/attachments/20160810/a7e83c97/attachment.html>