On Thursday 12 January 2006 13:36, Tomas Vilda wrote:
>
> we are going to buy two spanned card with HFC-E1 chips and we are
> interested in E1 passive monitoring as described i Aculab's company page :
> http://www.aculab.com/products/dna/passive_monitoring.htm (diagram at the
> end of the page)
>
> can we use vISDN drivers for doing passive monitoring and is this driver
> capable of doing this?

There are several requisites to fullfill in order to be able to do that kind 
of monitoring.

Electrical:

You must be able to disable line termination on the card. Several HFC-E1 based 
cards have this possibility along with traditional 75/120 Ohm termination.

You should also create a Y-cable, if the cable is overall short you could do 
some brutal connection, a la 10Base2 if using 75 Ohm unbalanced E1s or with 
the solder if using 120 Ohm balanced E1 :)

Hardware:

You must be able to read both directions at the same time. The only way to do 
this with HFC-E1 is to use two chips (two boards are ok) connecting one 
direction to Chip1/RX and the other direction to Chip2/RX.

Drivers:

I did not write the HFC-E1 driver yet. It will be a matter of few days from 
the time I start, however I currently have other priorities. I do not plan to 
provide a HFC-E1 driver within one month.

The rest of the infrastructure is able to handle this kind of exotic 
connections, you will end up with two netdev interfaces and each one will 
receive frames going in each direction.

Layer2/3

This is where some work should be put. My L2/L3 implementations are of course 
designed as interfaces between upper layers and lower layers, a request from 
L4 will be passed to L3 which in turn will pass it to L2 and so on.

What you want to do, instead, is an implementation that would track L2 and L3 
just by monitoring what two other parties are transmitting to each other.

It's a completely different model, you would need to implement this kind of 
monitor nearly from scratch. (of course, you may recycle the message 
encoding/decoding part in libq931 and lapd modules)

Recording the calls payload would be easier as you would just need to open two 
streamports and connect them to the same B channel on both direction, read 
and write to file.

Bye,

-- 
  Daniele Orlandi
_______________________________________________
Visdn-hackers mailing list
[email protected]
https://mailman.uli.it/mailman/listinfo/visdn-hackers

Reply via email to