Small/medium business server platform
nejc at skoberne.net
Mon Jun 16 18:42:42 UTC 2008
I am a final year student of Faculty of computer and information science from Slovenia.
I've also been a Linux/BSD system/network administrator for quite some years now.
I am starting to write my thesis and I am looking for good ideas. I am mostly
interested in small/medium business system administration. I also would like to continue
the work which I would make a plan for, after the thesis.
We all know "Windows Small Business Server" platform, which provides many "roles" (as
MS calls them) which can satisfy many of the small company's needs (mail, file, DNS, AD,
print, web, ... server). But since I don't do Windows, I am interested in an equivalent
open solution. So far, I have found ClarkConnect and eBox. Both are based on Linux.
ClarkConnect looks promising, but is not completely open/free and I don't like that.
eBox, on the other hand, doesn't have such limitations, but seems quite "unfinished"
although it has been in development for 4 years now. Also, it misses a key feature I
would like to have.
Also, I know pfSense. I love it - also, because it's FreeBSD based and I really like
FreeBSD. But, for now, it is non-modular and doesn't provide other services besides
being a firewall/router/network appliance.
So I was thinking to start a new project, which would have these features:
- it is based on FreeBSD;
- it is monolithic (as pfSense - no third party modules, at least in the beginning)
but still modular (so you can turn on or off various modules you (don't) need);
- it provides a "configuration API" - this is the key feature I mentioned before:
pfSense too, only has a web interface, which is OK, because that's normally all
you need. But I think that it would be really nice if anyone could create his or
her own interface for managing configuration - many companies have many custom
application frameworks (at least the one I work in, has one) and providing this API
could make management of such servers very scalable. For example, it would be very
handy for me, because now I have 50 small servers at various customers and their
administration can be time consuming. Also, I am not the only administrator, neither
I want to do boring routine tasks (that's why I would need a unified GUI, so other
non-console people can do it).
- it is open, under one of the open-source licences, professional support could be
eventually charged to keep the project going
- the services are running in FreeBSD jails for maximum security, 1 service/jail
- it would offer features such: [most of the pfSense features], mail (POP3, IMAP,
SMTP, webmail, ...), DNS, file/printer services, http, ftp, trac/svn, fax, DNS,
... - the software for these functions is already out there - it's just that
someone have to put it all together
1. Is such a product needed? Would anybody use it? How to find this out?
2. Connected to the previous question: could such a project live if it was good?
I don't have experience in "developing OS software for money", but I also need
to make a living.
3. I would like to make an international team of developers eventually, like other
open-source projects have;
3. Which features do you find attractive/unnecessary/stupid and which do you miss?
4. If you were I and could start with something "new" as I am thinking about to,
what it would be? :)
Thanks a lot for your thoughts.
More information about the freebsd-questions