[Freedombox-discuss] A software architecture for the FreedomBox

[Rob van der Hoeven]

On Wed, 2011-04-13 at 18:20 +0200, Jonas Smedegaard wrote:
> On 11-04-13 at 05:49pm, Eugen Leitl wrote:
> > On Wed, Apr 13, 2011 at 05:06:23PM +0200, Rob van der Hoeven wrote:
> > 
> > > One of the first things you have to do when building a system like 
> > > the FreedomBox is figure out the software architecture. I have come 
> > > up with an interesting architecture that is made of Linux Containers 
> > > (Virtual
> > 
> > Using Linux vserver guests for service separation (jails on steroids) 
> > is an excellent idea, actually.
> 
> I find it interesting, but am sceptical: I believe each jail consume 
> separate memory for their libraries - i.e. cannot benefit from shared 
> libraries.  So I worry about memory consumption.
> 

I was worried about the extra memory requirements too. Still i really
wanted to have isolation of services so i decided to do some
experiments. From my pstree/free results you can see that i'm running 4
VM's and a bunch of other programs. My system only uses 190 MB leaving
326 MB free. My system is very badly configured (too much daemons
running), but still there is no reason to worry about memory usage. I
expect that future FreedomBox hardware has a minimum of 1 GB memory
making memory usage even less important. 

> On a related note I want to keep low the number of used library 
> environments - e.g. try to aim for...
> 
>   * only one of libssl or gnutls
>   * fewest possible of Python, Perl, Bash, Ruby, PHP
> 

One of the reasons i want service isolation so badly is to enable
services to use their own libraries. If you force a program to use a
particular version of a library it has to be built and tested with that
library. This delays the deployment or may prevent the deployment of a
service. Important parts of the FreedomBox are still experimental or
even non-existent. My FreedomBox architecture must be able to handle
this.
 
Rob.