[134306] in North American Network Operators' Group
Re: RES: RES: Software For Telcos
daemon@ATHENA.MIT.EDU (Daryl G. Jurbala)
Tue Jan 4 10:05:06 2011
From: "Daryl G. Jurbala" <daryl@introspect.net>
In-Reply-To: <E723F40B750C0D47B43CDC67EFC3806E1CE4CF@MAILSRV4.aquarius.cpqd.com.br>
Date: Tue, 4 Jan 2011 10:04:19 -0500
To: Takashi Tome <takashi@cpqd.com.br>
Cc: NANOG Operators' Group <nanog@nanog.org>
Errors-To: nanog-bounces+nanog.discuss=bloom-picayune.mit.edu@nanog.org
On Jan 4, 2011, at 9:04 AM, Takashi Tome wrote:
[snip]
> Put in other words, software knowledge is not enough, you must have a =
deep understanding of that business and the history of the system =
itself...
[snip]
This is the case 100% of the time, regardless of how many "top" =
developers/coders think otherwise. Regardless of market segment.
The last "top telco" I dealt with enough to get a feeling for their =
systems was using PeopleSoft and a gaggle of the things that come along =
with it: $200/hr PeopleSoft consultants. They also had a walled off =
fiefdom in engineering that no one else could touch except for maybe a =
few select people in the NOC using Rational Rose which contained all of =
the engineering docs and much of the information the NOC really needed =
to troubleshoot anything of substance (I'm remembering an incident where =
I had a circuit down from them with no light on my end, yet the NOC kept =
arguing that they had a link on their end......turned out they were =
looking at their copper port and didn't realize it went through some =
other box, which has completely unmonitored ports, to turn it into =
single mode fiber to send across the city to me and only engineering had =
the documentation to show this).
I'm not saying that I'd be the right person to even make the initial =
design document for a large telco management system, but I can tell you =
that once you've seen how it's currently being done you'll realize that =
many of them don't appear to have the right person either.
Just in my small (meaning millions of minutes a day, not 10s of =
millions) voip business, we've not been able to find much off the shelf =
software at any price that would help with much of anything short of =
your standard generic business type apps. We're using a combination of =
open source packages, some lightly modified, and some internally =
developed software. Its not optimal, and I think it would break at even =
5x our current head count, but there isn't enough of a business case to =
go to some roll-your-own-with-consultants-base-app like PeopleSoft or =
SAP.=
