[AusNOG] Off Topic - Brisbane recruitment recommendations

[Ben Buxton]

Given the statements made in this thread, ranging from mildly upsetting to
somewhat infuriating, I cannot resist.

First, automation:

On Sat, May 14, 2016 at 2:59 PM Chad Kelly <chad at cpkws.com.au> wrote:

>
> In terms of automation I've been using a lot more of it in recent years,
> but you still need to manually check things as sometimes stuff goes wrong.
>

Then your "automation" is not automation.

You need to close the loop, otherwise you're just scripting. Scripting !=
automation (it's really just beginner scale).

Much "automation" today is a bit like the web in the late 90's, or a lot of
SDN "solutions" - lots of buzzwords and claims of solving world hunger, but
there's an awful lot of hacks put together that just address a small part
of the overall picture.
Then when it all goes titsup people whinge about how bad an idea it was and
it never was going to work anyway.

Back when the dotcom bubble popped in ~2000, this is exactly what happened
- we entered the dark ages of the web for a while. Until properly designed
companies/products emerged.

Close the loop, people.

Tools like GTMetrics have come in handy as well for checking things like
> the loading times of websites as website speed has become increasingly
> important.
>

"Tools" (power or otherwise) have no place in automated infrastructure
except during those (hopefully rare) times when something really does go
amiss.
Think of the electricity grid - electricians have tools like multimeters
and the like, but the grid uses fully integrated sensors.

Next, specialist vs generalist:

"Specialists" have one place - contract roles. You want a monkey to push
the right buttons on your AWS console, etc, then this is the place for
them. As soon as your work deviates from this, they're outta there. Or
sticking around whinging about change.

Permanent roles need generalists. I don't want to rehash much, but
technology and your infrastructure will change over time (and if not, look
up "Kodak", or "Novell", or...) - You need someone who will understand the
principles and happily change to the new dashboard with latest shiny
buttons that deep enough down do the same things. Not someone who will
requiring having a "crucial conversation" with about dealing poorly with
change.

Given a choice between an engineer who can explain how to use a vendor's X
and one who can't, but can explain how/why a generic X might work
internally, I'd much rather take the latter.

And then we come to pay...

If you're finding no candidates for a 100k role asking networking, "cloud",
automation and the like - the problem is not lack of candidates, it's that
no one wants to accept a dismal rate of pay, particularly in a capital.
You'll probably get a "conf t" engineer at that rate, but not someone who
can close the loop and tie all your infrastructure together.
If you have engineers now who have these skills and they're topping out at
100k, then you've done a great sell on the job and/or they're just
misinformed on the true rate out there - it's a great disservice to them.

Network engineers are going the way of the car assembly line worker. The
old jobs are going away, and most of the workers wielding spanners will be
made redundant. But some adapt and shift to designing the robots on the
line. Those few who take the latter path command higher salaries, and
rightly so.

BB
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ausnog.net/pipermail/ausnog/attachments/20160514/e73b7c1a/attachment.html>