Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!accuvax.nwu.edu!nucsrl!telecom-request From: dave o'leary Newsgroups: comp.dcom.telecom Subject: Re: AMI on T1 Lines Message-ID: <12512@accuvax.nwu.edu> Date: 24 Sep 90 02:44:06 GMT Sender: news@accuvax.nwu.edu Organization: TELECOM Digest Lines: 70 Approved: Telecom@eecs.nwu.edu X-Submissions-To: telecom@eecs.nwu.edu X-Administrivia-To: telecom-request@eecs.nwu.edu X-Telecom-Digest: Volume 10, Issue 672, Message 10 of 10 Roger Fajman wrote: >I would appreciate a short definition of exactly how AMI works on T1 >lines. I took a short course from Datatech Institute recently on >T1/T3 technology. (I liked the course, by the way.) Now I would like >to compare AMI to B8ZS for meeting one's density requirements on our >clear channel T1s, but can find only the definition of B8ZS in the >course notes. AMI is mentioned as being less preferable, but is not >defined. AMI is Alternate Mark Inversion. Basically it means that each ones bit is opposite in polarity from the ones bit proceeding it - this is used to deal with capacitive effects, etc. If two consecutive ones bits are received on the line with the same polarity, it is called a bipolar violation. These could be due to noise or a bunch of other factors. One's density is defined in different ways, with a certain number of ones bits necessary to maintain line energy and keep the repeaters in sync. FCC Part 68 and the AT&T 62411 specs disagree on what ones density means. I can provide more details if anyone cares. AMI does nothing to meet one's density requirements by itself - it just takes a synchronous bit stream and flips the ones bits. B8ZS is a technique used to maintain ones density - when the bit stream contains 8 consecutive zeroes, the "Binary 8 Zero Substitution" code is inserted in place of the 8 zeroes. This code intentionally contains a bipolar violation. This is why you have to work it out with the telco when you want to run B8ZS - their repeaters have to pass the bipolar violations (BPV's) rather than "fixing" them. Another technique for meeting ones density is called bit stuffing, i.e. you simply clock the DTE slower than 1.536 Mb/s and but a one bit into each byte, which is stripped out at the other end. When you work out the arithmetic it comes out to a DTE bit rate of 1.344 M/s, your ones density is insured, and AMI works as usual. The bit stuffing is not standardized between CSU's :-( so if you want different CSU manufacturer's equipment to talk to each other on different ends of the T1 (and not lose the extra bandwidth) then you need to run B8ZS. However, not all phone company equipment does B8ZS, although this seems to be getting a lot better. I guess this is why they say that B8ZS is preferable. Our C&P sales guy told us that B8ZS costs more and that we need to run ESF to use it (which I didn't understand...if anyone can explain that one I'd appreciate it). So basically you are always kind of running AMI, its just that B8ZS allows the special BPV's to get through. >A reference to look it up in would be helpful too, especially if it's >a publication likely to be found in a computer-oriented library. I've seen a bunch of books on the digital hierarchy around - however the ones I've purchased aren't with me right now. Two that I would recommend are a book by Bernard Keisler, which I can't remember the title of (something like Digital Transmission Systems, it is yellow with black lettering) and another book called something like "Megabit Communications Systems", I can't remember the author (it is orange and black). The Keisler book is older and is more technical/mathematical. The other book is new, from this year I think, and covers a wider range of material. I should have these books back by later this week, let me know if you can't find them. I saw the "Megabit.." book at the Maryland Book Exchange and at Reiter's downtown today. Good luck...drop me a line or give me a call if you have other questions... dave o'leary oleary@noc.sura.net SURAnet NOC Mgr. (301)982-3214