[Asterisk-Dev] IBM/SGI implementations

John Todd jtodd at loligo.com
Sat Nov 20 14:28:48 MST 2004


[new thread started]

At 3:07 PM -0600 on 11/20/04, Steven Critchfield wrote:
>On Sat, 2004-11-20 at 14:42 -0500, John Todd wrote:
>
>>     IBM: I'm waiting.  Step up to the plate, guys - this, along with
>>  many other Asterisk improvements, is where you could make a
>>  difference in telco-grade Linux.  Same with OSDL.org - there seems to
>>  be a no focus on applications, despite that being where the greatest
>>  changes could happen for market uptake.
>
>After a presentation by SGI to our LUG, it was mentioned how some of
>their Altix hardware would possibly be a good fit for asterisk as well.
>The Altix hardware allows you to add CPUs and memory to the problem as
>seperate entities of the cluster. Also the P-bricks may allow some
>people the ability to use more than 2 quads at once.
>
>I would love to see SGI and/or IBM jumping in to see how easy it is to
>kit up a really nice telecom system on their hardware.
>--
>Steven Critchfield <critch at basesys.com>

I agree; the more the merrier.  There was commentary in a private 
discussion about the IBM Power5 eServer OpenPower 720 platform, which 
can be configured in a 4-way design with SUSE Linux (perhaps also 
YDL?)  Getting Zap drivers to work with a DS3 card in this 
architecture might even result in a do-it-all platform that can even 
transcode 672 channels into a high-complexity codec.  I haven't the 
slightest idea if it will work or not, or if it's "better" than 
competing chipset/vendor implementations, but it looks promising and 
I'm hoping that someone might have news of this as a follow-up to 
this thread (you know who you are.)

I hadn't thought about SGI.  Do they have any special hardware tricks 
up their sleeves for perhaps doing codec transcoding in a more 
efficient manner than in the "generic" main CPU?

Of course, the trick (as Race noted in the introduction to his DS3 
thread) is getting the following components to the right place at the 
right time for any new implementation:

   - a demonstration platform (supplied by the vendor or VAR)
   - a very competent Linux-oriented C coder who is familiar with Asterisk
   - time
   - motivation (in the form of money, glory, fame, whatever)

JT



More information about the asterisk-dev mailing list