[flow-tools] New flow-stat design
Horatio B. Bogbindero
wyy@admu.edu.ph
Thu, 13 Jun 2002 15:12:27 +0800
--EeQfGwPcQSOJBaQU
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable
On Thu, Jun 13, 2002 at 02:55:14AM -0400, Mark Fullmer wrote (wyy sez):
> On Thu, Jun 13, 2002 at 08:59:04AM +0800, Horatio B. Bogbindero wrote:
> > Surlignage Mark Fullmer <maf@splintered.net>:
> >=20
> > > (I'm resending this, pair appears to be dropping mail yet again)
> > >=20
> > > I'm working on a redesigned flow-stat. New features:
> > >=20
> > will the old flow-stat still be maintained at least for a few more rele=
ases?
>=20
> Yes, same with the old flow-filter. The new flow-nfilter needs time to
> mature, it's had no time in the field yet.
>
don't worry. i'll give it sometime in the field. abiet limited to i386 syst=
ems.
hehehe.
=20
> > just as a suggestion. how about an XML formatted output with its=20
> > corresponding DTD for flow-tools. abiet slow and crufty, it could provi=
de=20
> > greater flexibility later on when XML catches on. just an idea.=20
>=20
> The idea is have flow-stat do the number crunching and export a simple
> to parse format then have other front ends written in perl or python
> which produce ASCII, HTML, RRD, etc.
>=20
i see. must be some generic and easily parsable format i support. of
course, wouldn't it be cool to at least have an HTML frontend part of
the distribution already. (as you said probably in python or perl)
> > interesting and nice. good luck! just a question will this be out befor=
e=20
> > version 1.0?
>=20
> The new flow-stat will be in 0.59 if there aren't any major bugs in
> 0.58. Probably a few months depending on how distracted I am with
> other projects.
>=20
that cool. the new flow-stat will save me a great deal of piping. basically,
i have a PHP frontend that simply calls that series of piped commands
and outputs them onto the HTML page. with the new flow-stat it would be much
cleaner.
> What feature do you (or others) want to see for a 1.0?
>=20
-95th percentile billing seems to be a good thing
-ability to breakdown output into shifts (8:00-6:00 and 6:00-8:00=20
for example). this could probably be done with a time start and
time end parameter to the flow-nstat config. i do this using a shell=20
script and piping the commands and using flow-iptime (the crufty=20
little thing migz and i wrote). which as you can see is really quite=20
messy. we have this feature for billing since early morning hours
are much cheaper than primetime hours.
hehehe.
anyway, when i think of something i will be sure to let you know. keep
up the good work.
-------------------------------------------
William Emmanuel S. Yu
Ateneo Campus Network Group (AteneoCNG)
email : wyy at admu dot edu dot ph
web : http://CNG.ateneo.net/wyu/
phone : +63(2)4266001-4186
GPG : http://CNG.ateneo.net/wyu/wyy.pgp
=20
War spares not the brave, but the cowardly.
-- Anacreon
=20
--EeQfGwPcQSOJBaQU
Content-Type: application/pgp-signature
Content-Disposition: inline
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org
iD8DBQE9CEXbOgIOlr0CsAERAoF4AKCu7oYrUxbfJvHr0PtmWvU+sNL2WwCfer2Q
fezQYNgPfCKnu4ooj5jtR6E=
=2il0
-----END PGP SIGNATURE-----
--EeQfGwPcQSOJBaQU--