home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!spool.mu.edu!yale.edu!ira.uka.de!smurf.sub.org!ppcnet!sks!mcd!mips.ruessel.sub.org!naddy
- From: naddy@mips.ruessel.sub.org (Christian Weisgerber)
- Newsgroups: comp.dcom.isdn
- Message-ID: <H.eg.37&NepFo9TQ@mips.ruessel.sub.org>
- Organization: My Individual Private Site
- Subject: Re: Asynchronous TAs
- Keywords: RS-232 asynch synch 57.6K 56K
- Summary: where are they?
- References: <1992Dec23.192710.3006@xenitec.on.ca>
- Reply-To: naddy@mips.ruessel.sub.org
- X-Software: HERMES GUS 1.10 Rev. Dec 18 1992
- Date: Fri, 25 Dec 1992 23:26:42 CET
- MIME-Version: 1.0
- Content-Type: text/plain; charset=US-ASCII
- Content-Transfer-Encoding: 7bit
- Lines: 74
-
- In <1992Dec23.192710.3006@xenitec.on.ca>, Vance Shipley writes:
-
- > Of all the ISDN terminal adaptors I have seen very few provide asynchronous
- > speeds above 19.2K. None of the TAs I have seen provide faster than 38.4K.
-
- See below.
-
- [...]
- > There is a popular misconception that asynch can only support 19.2K. This
- > is simply untrue, any lowly 16540 UART can handle 57.6, 76.8 and even 115.2K!
-
- Yes and no. The electrical specifications for a standard serial port,
- V.28 (->EIA232), set a maximum transmission speed of 20000bit/s. That's
- also the reason that V.110 is limited to 19.2kbps. The CCITT can hardly
- disregard its own recommendations.
-
- BTW, if we take this approach, I would like to see it implemented this
- way:
- The connection DTE<->TA should run at a commonly supported speed close
- to or higher than 64kbps, with hardware flow control i.e. RTS/CTS, the
- TA should bidirectionally buffer the data, strip start/stop bits, and
- run a standardized HDLC-type protocol (LAPB?) over the ISDN line, This
- way we also have flow control DCE-DCE, and most important efficient
- usage of the B channel. Basically the TA should just behave like a
- modern high speed error correcting modem, only faster. Running V.42bis
- DTE<->DTE would be an obvious extension.
- I think this scheme is already included in CCITT recommendation V.120.
-
- > This means that almost all PCs in use today can handle the speeds I suggest.
- > The cost of doing synchronous on a PC today is astronomical in comparison.
-
- Actually I think TAs are the least interesting for PCs. The PC market is
- so huge that it can sustain simpler machine-dependend solutions. An
- external 57.6/76.8kbps TA would be more appealing for other
- architectures where the prices of machine specific ISDN devices
- skyrocket. Not all the world is an "industry standard PC"
-
- However, the overhead or "cost" of running high speed asynchronous
- serial connections is rather high. If an interrupt is signaled for every
- character you tie up a lot of CPU power. Even with buffered UARTs you
- don't want to do this with a Unix system, do you? You can avoid this by
- using intelligent I/O subsystems offering (multiple) serial ports
- managed by a micro-controller of its own, but that's not exactly cheap
- or simple either.
-
- So, why go through the async-serial bottleneck at all? Especially with
- PCs?
-
- > Of course the other choice here is to get a TA which goes in the PC and
- > eliminates the asynchronous problem. This is only a solution for DOS users
- > because of driver availability.
-
- The cheapest way to provide ISDN for a PC is a simple expansion card
- with no intelligence of its own. Two chips by Siemens (AMD), some low
- level stuff, that's it. About as difficult as a serial card. In quantity
- of a few thousand production price should be about $70 and will become
- cheaper when ISDN chip prices drop. Very competitive when compared to a
- high speed modem.
-
- Yes, most work has to be done in the driver or application layer. Yes,
- this can be a problem with the multitude of operating systems around.
- But not any more than with other high performance peripherals like hard
- disk controllers, video boards etc. However, multi-layered drivers
- should provide high portability between different platforms.
-
- Within a few years ISDN will be everywhere. In Germany Telekom has
- announced nationwide ISDN availability for the end of 1993. (We already
- have urban availability.) From what I read in this group the U.S. seems
- to be legging behind, but for how long?
-
- I think it's time for some serious industry standards concerning ISDN.
-
- --
- Christian "naddy" Weisgerber, Germany naddy@mips.ruessel.sub.org
-