I don't know if you are referring to a comparison of connect times between
the MAX and the PM2 for ISDN, or modem calls. If it was comparing modem
connect times, the modem attached to the PM2 has a major affect on time of
call connects. How well the modem handshakes, retrains, etc. Not all modems
are created equal.
> As far as performance issues, go, that isn't surprising at all, and I do
>not disagree with it. I haven't seen any performance problems here, but
>we don't run it too hard, and have pretty much written off ever using the
>other 2 PRIs for anything.
I believe one reason for the MAXs peformance bottleneck is its single
processor architecture. Even though it is a RISC, it has to handle all
operations for up to 96 calls. In contrast, the PM3 employs a
multi-processor architecture where the communication process is shared
between many processors. An example, in the PM3 the DSP on each modem card
is responsible for running the modem, compression and PPP framing processes.
This not only off loads the main CPU, but because the data doesn't have to
travel from the modem dsp to another processor (for PPP framing), it
eliminates unnecesary memory read / writes. That means near no latency, and
increased performance.
> I doubt that the PM3 could handle 46 B-channel PPP connections, all
>sending and receiving minimum sized IP packets either (which is the
>highest possible theortical load). I wish more manufactures would give
>their products a packet-per-second rating (I believe only Gandalf does
>this know, but their software stinks).
>
As the product manager for the PM3 I am exploring options for documenting
PM3 performance test results. I'll keep you posted. ;-)
--Marty
****************************************************************
Marty Likier Livingston Worldwide Headquarters
Product Marketing Mgr. 4464 Willow Road
(800) 458-9966 Pleasanton, CA 94588 USA
(510) 737-2164 (510) 244-1903 (FAX)
marty@livingston.com Web: www.livingston.com
****************************************************************