Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!decwrl!hayes.fai.alaska.edu!accuvax.nwu.edu!nucsrl!telecom-request From: BRUCE@ccavax.camb.com (Barton F. Bruce) Newsgroups: comp.dcom.telecom Subject: Re: AMI on T1 Lines Message-ID: <12583@accuvax.nwu.edu> Date: 24 Sep 90 22:58:35 GMT Sender: news@accuvax.nwu.edu Organization: Cambridge Computer Associates, Inc. Lines: 44 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 677, Message 6 of 9 In article <12480@accuvax.nwu.edu>, Roger Fajman writes: > 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 AMI simply means Alternate Mark Inversion, and translated to english means that each one bit will have the opposite polarity of the preceeding one regardless of how many zero bits are in between. In fact if a plus follows a plus or a minus follows a minus, that is a bipolar violation and is an error in a pure AMI system. The dumb repeaters need to see some ones to keep their clock in step, so long strings of zeroes are to be avoided. All B8ZS does is replace a string of zeroes with a recognisable pattern providing some needed ones that could not be normal data, and that is itself balanced around zero (same number of plus 1s as minus 1s so there is no net DC component). This recognisable pattern is simply replaced with the equivalent number of zeroes at the far end of the span, but the repeaters see the ones they need to keep their clocks in sync. The special pattern B8ZS uses has two bipolar violations. There will be a plus followed by a plus and a minus followed by a minus. The B8ZS notes you found probably show that special string being substituted for eight zeroes. If you simply DON'T do that substitution, you have a vanilla AMI line. A vanilla AMI line can't guarantee ones density if all bits are to be available for your random use. If you rob a bit fron each DS0 you get your 56kb rate rather than the 64 you should have. Some T 1/2 fractional services provided by a LEC may give you clear channel on 'your' 12 DS0s, but they can simply ram ones on in 'their' 12 DS0s. Theirs will be every other time slot. You get the odd and they get the even, or vice-versa. That way there can NEVER be long strings of zeroes no matter what you send, and they can use most of their OLD hardware to provide you the service. YUK. Clear as mud, right?