[Asterisk-Dev] asterisk 'stable'?

Greg Boehnlein damin at nacs.net
Tue Sep 13 17:11:06 MST 2005


On Tue, 13 Sep 2005, Brian K. West wrote:

> On 9/13/05 3:29 PM, "Kevin P. Fleming" <kpfleming at digium.com> wrote:
> 
> >> the asterisk 1.0 is named stable, and as i've been told that's  because
> >> it's feature-frozen.
> > 
> > And we've already decided to stop using that nomenclature for future
> > releases.
> 
> Yes that was a bad choice to call it "stable".
> 
> > 
> >> will there ever bee a STABLE asterisk, as in 'does not crash'?
> > 
> > What else do you have in your life that is perfect and never experiences
> > any problems? Of those things, how many of them are as configurable and
> > run on as wide a variety of systems as Asterisk does?
> 
> Lets see.  Apache, MySQL, OpenH323 and many more run without issues and run
> on a wide variety of systems.   So that argument is totally false.

Bullshit. I have issues that cause Apache and MySQL to crash under certain 
circumstances that can be easily replicated in normal production 
environments. Nothing out of the ordinary, but we know the limitations 
of the versions of the software that we are running and we work around 
them. No software is perfect, and no software is every completely crash 
free. Is Linux perfectly stable and crash free? No way. It depends on a 
lot of factors, such as hardware, drivers, release level, compiler, 
optimization settings, libraries etc.
 
> > Let's be a little realistic here, OK?
> 
> These are very realistic goals to have software that does what it should and
> does it without fail.  Asterisk is very routinely tested on low load systems
> with very few calls and these problems will never show up under those kinds
> of loads.  When you load a system up with 200 calls you'll start to see all
> kinds of weird things taking place that usually don't show up.  Asterisk
> should be tested with in the lab with these types of configurations.

Same goes for Apache and MySQL. Load them up under certain real-world 
situations and they will act in unexpected ways. Do the same thing with 
Windows, and you'll get even stranger results!

My opinion is that Asterisk, as it was originally designed, I.E. as a 
low-volume PBX replacement has done a fantastically stellar job of 
overchieving. Where it seems to be failing is where people are pushing it 
beyond the original intentions. Most asterisk implementations that I hear 
about are successful and stable. Granted most of those do not exceed 96 
channels on a PRI card, but for pete's sake.. if you want a 
true Softswitch, go buy a Sonus.

Asterisk is a VERY stable PBX replacement and low volume Gateway / 
Application server. When you start to scale it much beyond that, you are 
exceeding it's original design implications. As such, you are most likely 
going to encounter situations that were never even dreamed of when the 
code was birthed. That is what people are starting to do now, and we're 
seeing where Asterisk falls over. This a natural part of any development 
cycle. The beauty of Open Source development is that the entire world can 
take part in it and appreciate the value of the software, warts and all.

-- 
    Vice President of N2Net, a New Age Consulting Service, Inc. Company
         http://www.n2net.net Where everything clicks into place!
                             KP-216-121-ST






More information about the asterisk-dev mailing list