On Thursday 09 February 2006 08:43, Tony Davis wrote: > Hope this is not a stupid question, but I have just installed a T1 in my > office for my data services. I assume that this T1 is a PRI and is just > configured to be one big data pipe. From what I understand there are 23 > 'B' channels (and a signalling channel) that comprise the T1. Is it > possible to have, say, a few of the B channels configured as Voice > services lines with associated DID numbers and the rest of the bandwidth > for internet data? Or even better, can the number of voice channels be > dynamic based on the traffic?
I am almost certain that you have a CAS T1, not a PRI. A T1 has 24 8-bit channels and a framing bit that are sent 8000 times a second. For data, those 24 8-bit channels carry your traffic. For voice, though, each channel needs to be able to convey its state (on hook, off hook, ringing, busy, etc.) and that is where you have several options. I won't bore you with the details, but your two most common options are either CAS or CCS. CAS = Channel Associated Signaling, also known as RBS or Robbed Bit Signaling. Each channel has some of its bits "stolen" or "robbed" from it in order to convey channel state. You never notice the missing bits with traditional voice/analog data calls. However you are quite limited with what these voice circuits can do, and when you start needing things like caller ID, DNIS/RDNIS and so on things get hairy really fast. Just look at some of the past posts here to see. :-) CCS = Common Channel Signaling. The most common form of CCS for us mere end-users is PRI. Here you sacrifice one entire 8-bit channel for signaling. This is where your 23 Bearer (B) channels and 1 Data (D) channel come from when you hear people talk about 23B+D with PRI. You have 23 completely "clean" channels and one channel that is dedicated to nothing but tossing around channel state. However you are now conveying state in a very "rich" format. You can send arbitrary data and ISDN has many, many extensions to pass along all kinds of data, and there are all kinds of other issues you run against regarding very specific interpretations and implementations of the ISDN specification and its many sub-parts. If you are looking to "steal" some of those T1 channels for voice, you are most likely going to have to work with an add-drop DSU/CSU (does exactly what it sounds like it's doing) or plug it into a Linux box which is able to split off some channels for voice and others to the networking stack. (Asterisk supports this both with Digium cards and Sangoma cards.) Just be aware that I have never seen this done dynamically. You and the telco decide how many channels you get for voice and the remainder is for data, and what you choose is what you get. Every voice channel you want drops your data rate by 64kbps. It's possible to do it dynamically (i.e. your data rate rises and drops depending on the number of active voice calls), and I believe that the Sangoma drivers support this but I don't know of a single telco that allows it. Dynamic channel allocation is a great solution for connecting branch offices through a dedicated point-to-point T1 where you have full control over what is connected to both sides of the link. Hopefully I've answered your questions. :-) -A.
