Re: Thinking sideways...
- From: "Ron N." <rhnlogic@xxxxxxxxx>
- Date: Thu, 30 Aug 2007 14:07:58 -0700
On Aug 30, 10:34 am, Vladimir Vassilevsky <antispam_bo...@xxxxxxxxxxx>
wrote:
Iwo Mergler wrote:
Hi all,
I have an unusual problem and a feeling that there's got
to be a branch of mathematics or signal processing dealing
with this class of problems.
The problem is about fitting digital PCM sample streams
of different sample rates into a time-multiplexed bus.
[...]
I'm in search of an algorithm to automatically fit a given
set of sample streams into a given frame structure while
minimizing the overall required bus frequency.
To make things more interesting, each sample stream comes
with a jitter requirement - related to buffer sizes. In
other words, there must be a minimum number of blocks per frame.
I can't help but think that the problem has some similarities with
prime number related algorithms and also modulo arithmetics. That
is, the headache feels familiar.
The transport looks very similar to ATM.
Yes, and some classes of ATM implementations use
various rate control and QoS schemes to minimize
the buffering requirements. Used by the phone company
to multiplex lots of independent streams each with
some guaranteed data rate. Lots of literature on the
subject.
Many years ago, I helped architect a portion of an
ASIC reputedly used in the data paths of some of the
most widely deployed OC-12 ATM line cards.
IMHO. YMMV.
--
rhn A.T nicholson d.0.t C-o-M
.
- Follow-Ups:
- Re: Thinking sideways...
- From: Iwo Mergler
- Re: Thinking sideways...
- References:
- Thinking sideways...
- From: Iwo Mergler
- Re: Thinking sideways...
- From: Vladimir Vassilevsky
- Thinking sideways...
- Prev by Date: Re: Thinking sideways...
- Next by Date: Re: Professional Exams for Digital Signal Processing!?
- Previous by thread: Re: Thinking sideways...
- Next by thread: Re: Thinking sideways...
- Index(es):