home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.dcom.isdn
- Path: sparky!uunet!paladin.american.edu!darwin.sura.net!cos!cos!bob1
- From: bob1@cos.com (Bob Blackshaw)
- Subject: Re: BONDING and T1 Letter Ballot 317
- Message-ID: <bob1.727625771@cos>
- Organization: Corporation for Open Systems
- References: <Kevin.Kershaw.3@StPaul.NCR.COM> <1993Jan19.075543.16763W@lumina.edb.tih.no> <1993Jan19.135401.21097@cbfsb.cb.att.com> <1993Jan20.132319.10075W@lumina.edb.tih.no>
- Distribution: world
- Date: Thu, 21 Jan 1993 14:16:11 GMT
- Lines: 54
-
- In <1993Jan20.132319.10075W@lumina.edb.tih.no> ketil@edb.tih.no (Ketil Albertsen,TIH) writes:
-
- >In article <1993Jan19.135401.21097@cbfsb.cb.att.com>, deej@cbnewsf.cb.att.com
- >(david.g.lewis) writes:
-
- >> [...] There's a feeling among some in the industry, however,
- >>that (a) for a BRI, 2B is a useful bandwidth that is not supported by any H
- >>channel, and (b) limiting to 384 and 1536 kb/s is insufficiently flexible
- >>for user applications, and allowing NxDS0 where N is any integer from 2 to
- >>24 or 30/31 (on PRI, at least) is a useful and necessary capability.
-
- >In the bluebooks, there IS a 2*64 kbit option (table 4-6 and 4-8/Q.931).
- >Yes, I know that there is a difference between 1*128 and 2*64, but for many
- >purposes the existing standard would be good enough, and the justification
- >for inventing a different wheel rather weak.
-
- >Disclaimer: I do not know anything about the '92 edition, so things may
- >have changed there. Further, I am not sufficiently deep into the lower
- >layers to see all the consequences of the "Restricted Differential Time
- >Delay" (which allows the two 64 kbit streams to be out of synchronization
- >by a certain amount).
-
- I'm always tempted to call this R2D2 :-), however, not being in the
- meetings (I got out of that back in 87), a look at the S/T Layer 1
- frame structure leads me to believe that the ~50 ms factor in RDTD
- is simply a result of the separation of the B-channel octets in
- each frame. You will see that the greatest separation occurs between
- the first pair of B-channel octets and amounts to 5 bits. Now a 48
- bit frame at 125 ms approximates a bit time of 5 ms, ergo 5x5=25.
- Given a frame at either end of any connection 2x25=50. At least
- that's how I believe it works.
-
- Now, whatever type of state machine you are using at layer 2 should
- be able to accomodate this slight out of sync condition with some
- form of simple buffering (or elastic store, if you prefer that term).
- If I'm wrong will someone who was at the meetings please correct me.
-
- >Yet, many wheels have been invented simply because the inventor overlooked
- >existing solutions. Basing a 2B solution on existing standards may be a
- >good idea. I don't think it is essential to provide 3B, 4B, 5B,...29B -
- >after all, we manage with modem speeds in doubling steps, and with TDM
- >channels in quadrupling steps (in Europe - US is more irregular), with
- >noone claiming that eg. a 7200 baud modem standard would "increase
- >flexibility", now that we got 4800 and 9600.
-
- Also, this thread seems to have overlooked the H10 rate here in NA.
- H10 allows use of 22 channels just in case you only have a single
- PRI. If you used H11 on a single PRI you would kill your D-channel
- and would then be in a bit of trouble.
-
- Glad to hear you are a convert.
-
- Bob.
-
-