2 BCMs stacked for capacity

[ne57301]

We've had our BCM 3.6 for almost 2 months and it's been less than flawless. At this poinnt we're experiencing clipping, delay, what appears to be half-duplexing, and we are limited in where we can build modems.

Here's the configuration: we're a little too big to fit in 1 BCM with an expansion cabinet so our vendor installed 2 and connected them via IP. BCM 1, our Main BCM, has the expansion cabinet and all the analog and digital cards. BCM 2 our Gateway BCM, has our PRI, LDT1, and 8 copper trunks. Thus all our voice traffic passes over the IP connection. We found that the IP delay causes us some problems including (but not limited to) not being able to use modems. We put an analog card in the Gateway and put the modems on it and they work.

My question is - why not have the street circuits in the same BCM as the expansion cabinet so I can have as many phones as possible not experiencing the IP delay?

 

[jdavid99]

with all that ip traffic make sure your vendor added MSPEC III's and that "ALL" network devices are capable and programmed for QOS. Your data packets MUST be prioratized AND packetized in all equipment including any swithches. As for the modem problem make sure your clocking is set to Primary and not Master. The default is Master and it screws up modems and fax machines. Just a thought....

 

[luv2ski]

there is almost always delay from ip to ip..this is per nortel...why not ask vendor who installed why the configured that way....also why would you stack 2 bcms to fit you company...not quite what it was designed for

ski season has started

 

[NARSBARS]

jdavid99's ideas is very important. Along with the MSPEC IIIs make sure system memory was maxed for both systems.

NARSBARS

 

[ne57301]

Thanks for the ideas folks. I agree that the BCM isn't quite the correct soloution but it's the one the sales departments (mine and my vendor's) worked out with no input from IT, so I gotta make it work.

Do you seriously think more memory will help?

I am pressuring my vendor on this configuration. They say they configured this way because we wouldn't be able to transfer calls between the BCM's if the DNs were spread across 2 boxes (or at least we don't have 'anti tromboning' software to release the 2 trunks that could be in use). I think it was just easier for them to configure call routing. By the way, the system is isolated on it's own network, not involved in the data network at all.