[ previous ] [ next ] [ threads ]
 
 From:  "D. Ubevidste" <detubevidste at gmail dot com>
 To:  m0n0wall dash dev at lists dot m0n0 dot ch
 Subject:  [m0n0wall-dev] Re: [m0n0wall] The future
 Date:  Thu, 13 Oct 2005 15:04:43 -0400
OS Choice
----------------

1 Performance advantages benefit everyone, although those (of us) at the
bottom of the CPU scale most. If we still see m0n0wall as serving the
"fanless/no moving parts" market, then care should be taken not to trade
performance for other benefits.

Example: We have a soekris 4800 handling a 1.5M Wan link with several IPSEC
tunnels, and several segmented internal networks. It is trivial to pegs the
CPU moving files between those networks--it is just barely adequate, because
we don't traffic much between these networks.

2 Underlying stability is key, and offered by all of the BSDs.

3. Feature availability will probably rank #2. Free has edge here for latest
version availability, then Open, with Net trailing some. But most will be
available on all three.

Thoughts:
* Since most all of the OpenBSD components desired are available ported on
other BSDs with better performance, consider seriously those other options
over Open.
* FreeBSD definitely holds major advantages, from peripheral support, to
developer familiarity, to large available featureset.
* Dragonfly is probably ideal, and probably too immature
* NetBSD has decent system performance, the very best network performance,
and portability strengths, plus claimed "good design". Lacks a good module
system. M0n0wall could be a significant contributor to this BSDs ongoing
efforts, and the developers might reciprocate with support.

Survey uses and limitations
----------------------------------------
A survey of current users will be helpful, for sure, but it
*Does not necessarily represent true breakdown of users (silent majority?
set-and-forget crowd?)
*The m0n0wall project itself may espouse values that are different, maybe
radically so, from the users who run it, and ultimately it is to those
values that the project must maintain alliegance.
*Excellent for driving feature sets, but maybe not for driving core market
identificacion.

=============
First and foremost:
=============

The m0n0wall project should settle on clear values and goals for what it
should do. Once those are made solid, we have a rationale for existence, a
basis upon which to make decisions, a standard against which to compare each
option (OS, core language, GUI subsystem, &c.), and a rock to stand on in
the hurricane (getting carried away now....)


du