In my opinion, anyone running large pools of PM3's without
running an alternative term server is making a serious mistake.
I wish I had made the USR TC my second term server.
While I havent run the TC solely in a town, it wouldnt
surprise me if it produced a similar percentage of people
who cant connect/stay connected. Im going to try it tho and
see if its as bad as the PM3. We estimated that 10-15% of
our users had problems with the pm3 due to modem code.
The configuration of it wasnt the nightmare everyone tried to
scare me into believing. If its slightly more expensive, thats ok,
I figure it just got several hundred complaining users off my
back. We were in serious danger of becoming known as the ISP with the
most disconnects/negotiation problems. We were the only ISP
in our area running pm3's.
Its too bad lucent refuses to allocate the resources necessary to fix
the modem code. I have a vision of the programmers who know
the code is bad, trying to convince the management beaurocracy to
spend the money to fix the problem, unsuccessfully. I suspect that
the 'testing' department of R&D is way too small. Does lucent
maintain an archive of common client modems out there for testing
purposes? Do they test their new firmware with this archive of
modems before posting it? Do they have relationships with
gateway and compaq to test the modems they put in their
systems? Some lucent guy in suit is deciding they make enough money
without caring. Lucent could be doing alot more for its customers,
they could maintain a public database of client modem information,
(with ATIx info) showing which subsets of client modems/firmware wont
work with which lucent firmware. The members of this group would
be more than willing to enter the ATIx codes of the problem modems
into a web database if it would help lucent R&D isolate and fix
the problems.
My theory on the disc/nego. problems is that there are
resource leaks in the code. We now have a script rebooting the pm3's
in the middle of the night, and things are going better. We would
have a town that was doing fine, all of sudden taking a rash of
disconnect/nego. complaints, rebooting seems to help.
BTW, I recently had a PM3 that began locking up,
and lucent tech support confirmed to me that they have an
issue with lockups, and they RMA'd my unit. By itself
that isnt so bad, but add to it that lucent seems to have a policy of
not admitting its problems in public, and you have another
strike against them.
BTW#2, on the line2 cards, avoid them. We had line2 cards that
wouldnt work with a T1, but the line1 port or cisco/adtran
would. I confirmed this on more than 1 T1 (all of which were frame
relay out of a 5E). Also, they dont work with ESF/B8ZS coming out of
a DMS10, you have to go AMI/D4.
So now my rant is done, and Im ready to be blasted by the
ex-lucent techs who still have stock in lucent, who will
defend it to the death no matter how many problems it has.
Go ahead and address the problems with the USR,play up your
problems with me, downplay the pm3 problems, harrass me for spelling
if you want, my customers and I are MUCH happier running USR
alongside our pm3s.
Jay Teutenberg
Advertisnet Internet Services
-
To unsubscribe, email 'majordomo@livingston.com' with
'unsubscribe portmaster-users' in the body of the message.
Searchable list archive: <URL:http://www.livingston.com/Tech/archive/>