[Asterisk-Dev] Is there a need for a DS3 channel/driver?

[Scott Laird]

On Nov 18, 2004, at 12:49 PM, alex at pilosoft.com wrote:

> On Thu, 18 Nov 2004, Mike M wrote:
>
>> Sometimes in the so-called unchannelized streams you are informed of
>> where the frame boundry is in the bitstream.  With that information 
>> one
>> can create soft-channelization.  Perhaps this is unreasonable for a 
>> DS3
>> however.
> Given that its a tough problem to drive current generation of quad-T1
> cards without frame slips, I think doing soft-ds3 stuff is crazy. It 
> could
> be possible given enough buffers...But I think its crazy ;)

Yeah, but how many of the frame slips would disappear if you added a 
couple extra ms worth of buffering on the card?  If we were willing to 
add a bit of extra latency on TDM circuits and add a few bucks to the 
cost of the cards, it seems like the frame slip problem would 
disappear, because we wouldn't require hard 1 ms interrupt latencies 
anymore.  I mean, modern PCs can clearly handle over 1 Gbps of 
bus-mastered data traffic, so there's no real reason to assume that 
they couldn't handle 45 Mbps of voice traffic *if* the interface 
between the computer and the card was reasonable and we didn't have to 
do much more then slap headers on buffers and ship them out.

I'm not convinced that the economics for DS3 cards makes a lot of 
sense, but I don't think the technology is completely out of our reach, 
either.


Scott