RE: [USR-TC] HiPer Access Modem Compatibility
The partial-page download you mention sounds like an MTU setting, possibly coming from your RADIUS server. Past experiences like this lead me to believe your old TC rack ignored the information, while the new one will certainly use the setting, and a wrong setting will cause exactly what you are describing. Check to see if your RADIUS is sending back something like "Framed-MTU=576", which seems to be a "common" setting on some RADIUS servers, but it will really mess up your dial-up users. - joel -----Original Message----- From: ISN Support Staff [mailto:support@isn.net] Sent: Wednesday, November 12, 2003 11:47 AM To: Discussion "relating to the 3Com/US Robotics Total Control modem systems. Subject: [USR-TC] HiPer Access Modem Compatibility Hello, We recently switched one of our dialing pools from an old Total Control 1000 system (CISCO AS5100 terminal server cards with USR Quad Modem cards running analog) to a HiPer system with digital lines and v90 support, and now we seem to be having a lot of compatibility problems. Even when we've had our end users disable 56K on their modem, some of them have a problem of getting connected but not being able to access the Internet reliably (they get the first few packets of data, then the browser just hangs, waiting for the rest.) I've noticed that this seems to happen mostly with AOpen FM56 modems, but they do seem to work fine on our other TC 1000 units (non-hiper: Dual T1's with 12 digital quad modems running v90) Does anyone know of any compatibility problems I should be aware of, or is there any suggestions as to a good init string/script to use in the Hiper systems to make them more friendly with non-USR modems? _______________________________________________ USR-TC mailing list USR-TC@mailman.xmission.com http://mailman.xmission.com/cgi-bin/mailman/listinfo/usr-tc
Hi Joel, Bingo! I had to bring a new radius server online at the same time we switched to the HiPer system (the old one wasn't sending back the response packets to the HiPer) and I forgot to remove the MTU line from the users file. Doh! Thanks for the tip. On Wed, 2003-11-12 at 13:51, Joel - Fox Computers wrote:
The partial-page download you mention sounds like an MTU setting, possibly coming from your RADIUS server. Past experiences like this lead me to believe your old TC rack ignored the information, while the new one will certainly use the setting, and a wrong setting will cause exactly what you are describing.
Check to see if your RADIUS is sending back something like "Framed-MTU=576", which seems to be a "common" setting on some RADIUS servers, but it will really mess up your dial-up users.
- joel
-----Original Message----- From: ISN Support Staff [mailto:support@isn.net] Sent: Wednesday, November 12, 2003 11:47 AM To: Discussion "relating to the 3Com/US Robotics Total Control modem systems. Subject: [USR-TC] HiPer Access Modem Compatibility
Hello,
We recently switched one of our dialing pools from an old Total Control 1000 system (CISCO AS5100 terminal server cards with USR Quad Modem cards running analog) to a HiPer system with digital lines and v90 support, and now we seem to be having a lot of compatibility problems. Even when we've had our end users disable 56K on their modem, some of them have a problem of getting connected but not being able to access the Internet reliably (they get the first few packets of data, then the browser just hangs, waiting for the rest.) I've noticed that this seems to happen mostly with AOpen FM56 modems, but they do seem to work fine on our other TC 1000 units (non-hiper: Dual T1's with 12 digital quad modems running v90)
Does anyone know of any compatibility problems I should be aware of, or is there any suggestions as to a good init string/script to use in the Hiper systems to make them more friendly with non-USR modems?
_______________________________________________ USR-TC mailing list USR-TC@mailman.xmission.com http://mailman.xmission.com/cgi-bin/mailman/listinfo/usr-tc
Hi *, On Wed, Nov 12, 2003 at 02:17:26PM -0400, ISN Support Staff wrote:
Hi Joel,
Bingo! I had to bring a new radius server online at the same time we switched to the HiPer system (the old one wasn't sending back the response packets to the HiPer) and I forgot to remove the MTU line from the users file. Doh! Thanks for the tip.
Uhm, despite this being solved... I'm experiencing slow connections and even drops with some "conexant" HXF-PCI-modems. I tried several combinations: TCS-4.5 w/ and w/o V.92 switched on ( if it matters anyway) TCS-4.7 w/ and w/o V.92 switched on each 1. orig US-Robotics 56k PCI flash installed ( <= V.90) and 2. conexant-generic driver published on their web-site ( <= V.92) Over 95% of all tries ended in transfer-"speeds" at 500-600cps. Sh...low, eh?! The best transfer-rate was achieved with: AT&F+MS=V34,1,75,28800,75,33600 which gave me a "V34plus" connection on the server-side. Rate ca. 3kB/s and stable. Any ideas VERY welcome, as this kind of beast is distributed widely here in Germany :-\ And of course I do NOT want to change much if any on the server-side. Thnx in @vance, Oliver. [...] -- Oliver.Francke@telefonica.de fon. +49-5246-80-1389 mob. +49-171-5597734 I used to have a sig, but I've stopped smoking.
Hi Oliver, Have you tried +MS=v90 on the client side? The +MS=v34 turns off all 56K features (which is why the modem only gets 3-3.5k/sec with that init) +MS=v90 should lock it at v90 and not try v92 (although I'm not 100% sure on that) which might work better. We have a lot of those Connexant modems here in Canada too (especially in Canada) and I've found that with the v90 drivers (haven't tried v92, since our TC's don't support it) they do work as well as any other software modem (4-4.5k/sec on binary downloads.) There might be a difference between the North American and European drivers though. On Wed, 2003-11-12 at 14:46, Oliver Francke wrote:
Hi *,
On Wed, Nov 12, 2003 at 02:17:26PM -0400, ISN Support Staff wrote:
Hi Joel,
Bingo! I had to bring a new radius server online at the same time we switched to the HiPer system (the old one wasn't sending back the response packets to the HiPer) and I forgot to remove the MTU line from the users file. Doh! Thanks for the tip.
Uhm, despite this being solved... I'm experiencing slow connections and even drops with some "conexant" HXF-PCI-modems. I tried several combinations:
TCS-4.5 w/ and w/o V.92 switched on ( if it matters anyway) TCS-4.7 w/ and w/o V.92 switched on
each 1. orig US-Robotics 56k PCI flash installed ( <= V.90) and 2. conexant-generic driver published on their web-site ( <= V.92)
Over 95% of all tries ended in transfer-"speeds" at 500-600cps.
Sh...low, eh?!
The best transfer-rate was achieved with:
AT&F+MS=V34,1,75,28800,75,33600
which gave me a "V34plus" connection on the server-side. Rate ca. 3kB/s and stable.
Any ideas VERY welcome, as this kind of beast is distributed widely here in Germany :-\
And of course I do NOT want to change much if any on the server-side.
Thnx in @vance,
Oliver.
[...]
Hi there... On Wed, Nov 12, 2003 at 03:15:17PM -0400, ISN Support Staff wrote:
Hi Oliver,
Have you tried +MS=v90 on the client side? The +MS=v34 turns off all 56K features (which is why the modem only gets 3-3.5k/sec with that
yep... true..., but even with a V.92 capable-software and locking V.90 things do NOT get better *sigh* Even just checked out the US-Robotics magic ControlCenter with no better luck...
init) +MS=v90 should lock it at v90 and not try v92 (although I'm not 100% sure on that) which might work better. We have a lot of those Connexant modems here in Canada too (especially in Canada) and I've found that with the v90 drivers (haven't tried v92, since our TC's don't support it) they do work as well as any other software modem (4-4.5k/sec on binary downloads.) There might be a difference between the North American and European drivers though.
... mhm... might be. Some other hint: I configure DSPs mostly per "factory-default", as we are serving for MANY people here and normally this works. I got complains from this customer for especially this and ELSA-MikroLink stuff as well... might be related?! Thnx again, Oliver.
On Wed, 2003-11-12 at 14:46, Oliver Francke wrote:
Hi *,
On Wed, Nov 12, 2003 at 02:17:26PM -0400, ISN Support Staff wrote:
Hi Joel,
Bingo! I had to bring a new radius server online at the same time we switched to the HiPer system (the old one wasn't sending back the response packets to the HiPer) and I forgot to remove the MTU line from the users file. Doh! Thanks for the tip.
Uhm, despite this being solved... I'm experiencing slow connections and even drops with some "conexant" HXF-PCI-modems. I tried several combinations:
TCS-4.5 w/ and w/o V.92 switched on ( if it matters anyway) TCS-4.7 w/ and w/o V.92 switched on
each 1. orig US-Robotics 56k PCI flash installed ( <= V.90) and 2. conexant-generic driver published on their web-site ( <= V.92)
Over 95% of all tries ended in transfer-"speeds" at 500-600cps.
Sh...low, eh?!
The best transfer-rate was achieved with:
AT&F+MS=V34,1,75,28800,75,33600
which gave me a "V34plus" connection on the server-side. Rate ca. 3kB/s and stable.
Any ideas VERY welcome, as this kind of beast is distributed widely here in Germany :-\
And of course I do NOT want to change much if any on the server-side.
Thnx in @vance,
Oliver.
[...]
-- Oliver.Francke@telefonica.de fon. +49-5246-80-1389 mob. +49-171-5597734 I used to have a sig, but I've stopped smoking.
I was purusing the knowledge base and I see a comment that a Hiper Arc that is doing nothing basically (not routing calls) can handle a max of 21 sessions and a Hiper Arc that is in service and routing calls can do 7 sessions max. My questions is what are they referring to when they say "sessions"?? Are they referring to the number of mpip clients that a mpip server can handle? The number of bundles, locallinks??? Can anyone shed a bit of light on the verbiage? Thanks in advance, Todd
participants (4)
-
ISN Support Staff -
Joel - Fox Computers -
Oliver Francke -
Todd Bertolozzi