$110,000 for Gated Source Code

So far, I've had a large number of questions about a statement I made here
in NANOG. The statement was made in passing, while I was looking for
sources for something that would run BGP4 and allow us to transition from
static single-homing to dynamic multi-homing. Rather than answer, yet many
more inquiries, I am posting the answer here in NANOG, to fore-stall
further such inquiries.

MHSC prefers open-source whenever we can get it and prefer to run
reference-standard code. This is why we run sendmail v8.8.8, qpopper, BIND,
and others. We are a Caldera VAR and our servers are built up from Caldera
Open Linux Standard edition, with about 30 add-ons <groan>. We *do* have
BRU, Netscape, and other commercial binaries that have been bought, but not
before attempting to find acceptable reference-standards. We have also
provided DewPoint/Caldera with input towards an Enterprise Server
distribution for Linux.

Anyway, we've applied for an ASN and ran into the BGP4 requirement. Ergo,
we were looking for the reference-standard BGP4 implementation, which is
GateD. We (I) was *very* surprised at the attitude exhibited at
<http://www.gated.org>. Very much anti-commercial. But, that does not
matter to me as I have $other$ things to worry about (Paying the rent
around here is one of them <sigh>). Ergo, having run *that* trail to
ground, I posted a query here in NANOG, that Dean Anderson, Shane Wright,
Craig Labovit, and others have answered.

I now have gated v3.5.9, for Linux, and am building it now. I'll probably
run into trouble, as I do with most things that don't have a configure
script, or have an otherwise non-standard build process (Why is it that
*only* the academic originated stuff is such a PITA? <sigh>).

Back to the point, we are a commercial technology R&D house that is also
bringing up an ISP (MHSC.NET). As such, since we actually have a
gawd-awful-scummy profit motive (why else form a corporation?), we do not
qualify for Academic membership in the Gated Consortium. The prices that I
quote here are for the level of membership which MHSC would have to sign-up
for in order to obtain what we really need. There are lower-orders of
membership, but if you look at the URLs you'll find that the minimum is
$10s of K-bucks$ for the first year, plus an annual maintentance agreement
of, at least, half of the first year premium (and we thought we had
profit-at-heart). Being a startup, we don't need such expense (it comes out
of *my* pocket, personally, and I ain't that rich).

Again, this is not a rant at the gated folks. I frankly wasn't much
surprised. I thank all those who helped me obtain the sources for
gated-3.5.9. BTW, Flemming did comment that since gated was NSF funded, it
may be inappropriate for such fees to be charged. In reallity, the message
clearly states that "start-ups need not apply". They just want to keep out
the riff-raff. I've long-ago quit getting emotional about these things and
have come to expect it from acadamia-types.

The URL is ;

The latest public release of gated is 3.5.9 and is available at:

ftp://gated.merit.edu/net-research/gated/gated-3-5-9.tar.gz

Lots of information regarding licenses, bug reports, and releases can be
found at http://www.gated.org/

Cheers,
Shane

Thank you, Shane.

So far, I've had a large number of questions about a statement I made here
in NANOG. The statement was made in passing, while I was looking for
sources for something that would run BGP4 and allow us to transition from
static single-homing to dynamic multi-homing. Rather than answer, yet many
more inquiries, I am posting the answer here in NANOG, to fore-stall
further such inquiries.

MHSC prefers open-source whenever we can get it and prefer to run
reference-standard code. This is why we run sendmail v8.8.8, qpopper, BIND,
and others.

I see no need for open source when it is quite obvious you do not
understand anything that you are compiling.

We are a Caldera VAR and our servers are built up from Caldera
Open Linux Standard edition, with about 30 add-ons <groan>. We *do* have
BRU, Netscape, and other commercial binaries that have been bought, but not
before attempting to find acceptable reference-standards. We have also
provided DewPoint/Caldera with input towards an Enterprise Server
distribution for Linux.

Linux.. the choice of a gnutered generation.

Anyway, we've applied for an ASN and ran into the BGP4 requirement. Ergo,
we were looking for the reference-standard BGP4 implementation, which is
GateD. We (I) was *very* surprised at the attitude exhibited at
<http://www.gated.org>. Very much anti-commercial. But, that does not
matter to me as I have $other$ things to worry about (Paying the rent
around here is one of them <sigh>). Ergo, having run *that* trail to
ground, I posted a query here in NANOG, that Dean Anderson, Shane Wright,
Craig Labovit, and others have answered.

Well, you can probably write your own BGP implementation in about
a month. This would cost you considerably less than $110k assuming
a conservative valuation of your time ($4.25-$5.00/hr).

I now have gated v3.5.9, for Linux, and am building it now. I'll probably
run into trouble, as I do with most things that don't have a configure
script, or have an otherwise non-standard build process (Why is it that
*only* the academic originated stuff is such a PITA? <sigh>).

LINUX#make
sh: make: command not found
LINUX#Ihatethesefuckingcomputersasdf3252r23t2g
sh: Ihatethesefuckingcomputersasdf3252r23t2g: command not found
LINUX#goddamn these academic types
etc..

So far, I've had a large number of questions about a statement I made here
in NANOG. The statement was made in passing, while I was looking for
sources for something that would run BGP4 and allow us to transition from
static single-homing to dynamic multi-homing. Rather than answer, yet many
more inquiries, I am posting the answer here in NANOG, to fore-stall
further such inquiries.

MHSC prefers open-source whenever we can get it and prefer to run
reference-standard code. This is why we run sendmail v8.8.8, qpopper, BIND,
and others.

I see no need for open source when it is quite obvious you do not
understand anything that you are compiling.

Before inserting your foot further into your insulting mouth, visit my
personal web-site. We also have a couple of PhD CS's around here somewhere,
oh yeah, they're working on paid-for commercial projects. How much money
have you stolen from the NSF to-date?

We are a Caldera VAR and our servers are built up from Caldera
Open Linux Standard edition, with about 30 add-ons <groan>. We *do* have
BRU, Netscape, and other commercial binaries that have been bought, but not
before attempting to find acceptable reference-standards. We have also
provided DewPoint/Caldera with input towards an Enterprise Server
distribution for Linux.

Linux.. the choice of a gnutered generation.

Bigotry and snobry, a sure sign of insecurity and penus envy. What's the
matter, didn't get laid last week-end?

Anyway, we've applied for an ASN and ran into the BGP4 requirement. Ergo,
we were looking for the reference-standard BGP4 implementation, which is
GateD. We (I) was *very* surprised at the attitude exhibited at
<http://www.gated.org>. Very much anti-commercial. But, that does not
matter to me as I have $other$ things to worry about (Paying the rent
around here is one of them <sigh>). Ergo, having run *that* trail to
ground, I posted a query here in NANOG, that Dean Anderson, Shane Wright,
Craig Labovit, and others have answered.

Well, you can probably write your own BGP implementation in about
a month. This would cost you considerably less than $110k assuming
a conservative valuation of your time ($4.25-$5.00/hr).

Buddy boy, you can't afford me <sheesh>!

I now have gated v3.5.9, for Linux, and am building it now. I'll probably
run into trouble, as I do with most things that don't have a configure
script, or have an otherwise non-standard build process (Why is it that
*only* the academic originated stuff is such a PITA? <sigh>).

LINUX#make
sh: make: command not found
LINUX#Ihatethesefuckingcomputersasdf3252r23t2g
sh: Ihatethesefuckingcomputersasdf3252r23t2g: command not found
LINUX#goddamn these academic types
etc..

Jeez, you can't even spell "damned" right. Also, in that context God is
capitalized.

Now, if Bradly wants to really play with something easy, install
Apache-SSL.mod-perl on a Linux box. Or anything else with a configure script.

<brad>

>I see no need for open source when it is quite obvious you do not
>understand anything that you are compiling.

<roeland>

How much money have you stolen from the NSF to-date?

<brad>

>Linux.. the choice of a gnutered generation.

<roeland>

Bigotry and snobry, a sure sign of insecurity and penus[sic] envy. What's the
matter, didn't get laid last week-end?

Can we dispense with the pleasantries and get back on-topic? :slight_smile:

It's unbelievable, the amount of e-mail I got from this. The intent was to
reduce, not enhance. For those that think this is off-topic to operations I
ask, "What is more pertinent to operations than a decision on which
implementation of BGP4 to run?" Yes, we are considering Cisco routers. Yes,
our ASN was approved by ARIN. Yes, Linux is indeed production ready, as far
as we're concerned. But, we're only a VAR, what do we know? Yes, we're a
conmmercial site. Yes, we are a GRS-site (Take that up with me in
DOMAIN-POLICY). Now, can we get to the point?

I suggest that those, who take these things personally, re-read what I have
paraphrased herein and let's get past the name-calling. The issue I'm
raising is, what to use for BGP4, in production.

Whatever your router supports. Talk to Ascend, Bay, Cisco, Jupiter and
Pluris and ask them about their backbone router products.

Did I say Jupiter!? Tsk, tsk.
I meant Juniper of course.