Jesper,
good to see you involved.
You comments make sense to me, and I'm glad to see that you brought your
excellent expertise on the mail topic to the table. Hence this
discussion list.
I suggest we take the SME Server functionality/role to a higher level,
and see what 'visions' are out there. And here is my take on things.
Today we see that most services (as in applications) are virtualized by
whatever platform. The aging OSI model may not apply to today's market
demands where both consumers and enterprise users want apps on their
phones, tablets and PC alike devices! (We've added another
virtualization layer). So your point on the core functionality (aka
Role) of SME Server is a valid one in my opinion. The core design dates
back to around the year 2000, and the world has progressed. Despite our
considerations on what 'we' consider to be 'our' target audience, things
have changed. Massively!
First of all, the above 'we' and 'our' may range from a home user to a
mid sized company and beyond (depending on the role of SME Server put in
play in that scenario), and requirements/wishes dramatically are
different. So the first question would be, 'who do we serve, what's our
role' ?
Make no mistake, we can make all things work, but why as a SME Server?
History shows that e-smith was developed to be simple, have
functionality and easily managed for especially non-profit organizations
that did not have dedicated IT staff. It should simply work out of the
box. Back in pre 2000 .....
14 years later, the core design has not changed (which is a good thing
regarding simplicity, stability and security). But just like you
experience, there are new demands out there. Especially after 14 years.
Again, today many IT staff do not care about the 'server' but care about
application provided functionality. (SaaS?) Hence they size up clusters
(even in small organizations, for it is easy to do. e.g. Proxmox) and
specialize in VMWare, Xen, Citrix, KVM and derivatives.
Easy deployment, scalebility and flexible management of functionality
for admins and app based end user experience is what drives the market
today.
Do we stick to the pre 2000 design and why, or do we look at tomorrow
and adapt? Either choice requires changes.
And yes, the SME Server development in my opinion is almost strictly
technically driven.Again this is a good thing re stability and security,
but waving off additional functionality to contribs can not continue.
SME Server must have a clear vision, functionality an a role in today's
market demands.
There are initiatives to make SME Sever more 'virtual aware' and
'container aware', but SME Server was never designed for that. CoreOS is.
So what is the role of SME Sever today and does it meet market demands?
I think we need to prevent the moment at one point in time that we find
ourselves as a hobby club carefully taking care of our old-timer, whilst
the rest of the world is driving their energy saving, subsidized,
hybrid, self guiding IA cars, with all the goodies and gadgets you want.
All controlled by your smart device. For sale at a bargain and very
affordable as trade-in programmes.
What would you do? I would by a next generation Tesla....
My 2 cents
-HF
On 09/21/2014 09:29 AM, Jesper Knudsen wrote:
I again think that this is approached the wrong way - most of these
suggestion are defined with technical reasons and may (I do not say I know
it) not reflect the use model of the SME servers where they are deployed.
Personally I think the roadmap should strengthen the functionality which is
used the most and which would affect most users/admin.
Am I wrong when stating that the SME server (core) functionality is:
- File server
- Web server
- Mail server
And all the rest is applied through contribs. We should try to understand
which contribs are most used and maybe integrate those in the core before
anything else. When it comes to the 3 core functions, then the web server is
OK now when PHP is upgraded to 5.3+ and when it comes to the file server,
then I am yet to see the great advantages of Samba4 for the SME server (its
primarily performance and encryption as I read it - not that security is not
important). The mail server though, has a hard time keeping up with the
increasing complexity of the spammers. Maybe some efforts could be made
here?
But all this is just my speculations, I think it's time to really figure out
how the thousands of deployed servers are used and improve their
functionality.
Just my 10 cents,
Jesper (aka Knuddi)
_______________________________________________
Discussion about project organisation and overall direction
To unsubscribe, e-mail [email protected]
Searchable archive at http://lists.contribs.org/mailman/public/discussion/
