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
