[Freeswitch-users] FreeSWITCH vs Yate

[Michael Jerris]

On Jul 22, 2008, at 9:39 PM, Mike Fedyk wrote:

> I've been using asterisk for a while now and just love running into  
> assorted deadlocks, crashes, quirks and SIP misimplementations (ie,  
> new Call-ID when forking/branching calls).  I'm currently looking  
> very closely at Freeswitch and another project that has been around  
> longer, Yate.
>
> Yate seems to have a good architecture, and looks like it may be a  
> good alternative to Freeswitch (if there was ever a need for it).
>
> Looking from the outside, it makes me wonder why Yate didn't just  
> get a few more developers instead of another project being started  
> instead.  The Freeswitch devs may have looked at it before starting  
> a new project and I'm curious as to why it was passed over.  Could  
> it be because Yate is written in C++?

I would say we tend to not be huge C++ fans, but we are not afraid of  
it by any means, we do use C++ in the places it has made sense for us  
and we do support modules fully written in C++.  I think the biggest  
difference between FreeSWITCH and Yate are in its approach to using  
other libraries and our licensing.  One of the main design goals we  
started with was to re-use as much already written code as possible.   
The best example of this is that we use the sofia-sip library  
heavily.  This library was 5+ years of development before we ever used  
it and is a massive piece of work.  Yate has often taken the opposite  
approach of writing nearly everything themselves.  I don't think this  
approach is sustainable for any project as you can not focus on  
everything and make it really good.  I spent several years watching  
the problems the asterisk project ran into and by far the most  
numerous were related to sip in some way.  I don't think any project  
that chooses the route of writing their own sip stack in the end will  
be able to handle the maintenance requirements without a massive user  
and developer base.  Another big difference would be that we wanted a  
license that was less viral to allow us to use non gpl-compatible  
libraries and to allow for a project that would be more interesting to  
businesses that are gpl averse.  As a result of this, I think we have  
been able to build a community that is much better supported by people  
and companies that are weary of working with gpl software or are  
looking to build custom commercial add-ons.  We have also spent quite  
a bit of effort to make the project very accessible to developers,  
allowing anybody the ability to create their own svn branches just by  
signing up and we try very hard to work with developers to make  
architectural suggestions and design ideas, with the idea of building  
a very vibrant developer community, the key in my mind to any open  
source project.  I think we have been fairly successful in this regard.


> Personally since even now Freeswitch has a bigger community I'd  
> probably only look at Yate to make sure I have a somewhat generic  
> event model for Vicidial (somewhere on my to-do list).

Not sure what you mean here.  FreeSWITCH has an event system, what do  
you mean by generic?

> This message is by no means meant to be taken as an attack on  
> Freeswitch.  I'm just curious.

No offense taken. :D

> Mike

Mike