[BC] Audio Streaming

[Alex Hartman]

IT guys who have been doing this a while know what a five-9's SLA is.
Some even still offer it. ;)

> However, I have seen a change in that mindset as I'm hearing more and
> more senior managers of non-broadcasting companies directing their IP
> departments to provide absolute continuous uptime and response to
> correct is a 24/7 mission.  No longer is 9-5 the uptime the objective
> and these young IP whippersnappers are getting a dose of real
> world/there is no such thing as a time clock demands on service.

I've carried a work cell/pager for years, been on-call for IT stuff,
etc. Now, in the IT world, finding the problem can be a little harder
than "oh, broken wire here". It can, and does take a little time to
find the real issue. Sometimes it's physical like bad drives, most
times it's PEBCAK. (Problem Exists Between Chair and Keyboard)

Now, any hosting company worth it's salt should know how an
infrastructure is built, and built correctly. Unfortunately these
companies are few and far between and out of the price range for many
"end users". Those that "get it" are people who design things like
www.thebunker.net and cyrusone.com. These types of datacenters
typically house the backup stuff for IBM, Microsoft, Apple, etc. I
know the bunker holds US and UK military servers as well.

Serious IT folk (engineers, administrators, etc) know how this works.
The 9-5 job is the support guy. If you pay enough, you get the bat
phone number to the pager/cell i'm carrying. But that's only if you
pay enough. A $24.95 a month customer is not going to get that phone
number or that level of service. Someone paying for a colocated
server, or a cage, or someone with a "high profile" will however get
the attention that their monthly invoice reflects.

That being said, having done IT engineering and administration, i have
my entire network monitored using various protocols, and using text
messaging i get messages of status updates, door openings, relay
triggers, etc. And if my servers become unresponsive, i have watchdog
SNMP power strips for that reason. If it doesn't come back up on it's
own, then i start to dig.

We also do care about the little guy, if it's effecting a large
number. Consider E-Mail services for a moment. If i have 40,000 users
on a monday morning checking their email, and the server crashes, yes,
you bet your ass i'm in there fixing it.

It's a totally different mentality between broadcast engineering and
beating the crap out of transmitters and replacing relays versus "find
the packet storm and stop it" or look, rogue perl script running away
and chewing all the CPU in a shared host, etc.

Moral of the story is, you only ever talk to the front line at these
companies, believe me, there's a LOT more behind the scenes going on
than anyone realizes. In other words, you're assuming that the support
guy is the one fixing the problem, when in reality, he's telling
someone who has the keys to fix it. But again, 90% of the time, it's
PEBCAK.

--
Alex Hartman