Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 23 Feb 96 12:54:45 +0100
From:      garyj@frt.dec.com
To:        peter%thirdeye.com@inet-gw-1.pa.dec.com
Cc:        freebsd-hardware%freebsd.org@inet-gw-1.pa.dec.com
Subject:   Re: ISDN Questions 
Message-ID:  <9602231154.AA00568@cssmuc.frt.dec.com>
In-Reply-To: Message from peter@thirdeye.com (Peter Rowell)  of Thu, 22 Feb 96  17:17:56 PST.

next in thread | raw e-mail | index | archive | help

peter@thirdeye.com writes:
> Hi,
> 
> A couple of questions about ISDN and FreeBsd.
> 
> 1.  What is the current status of the "II ISDN Interface"
>     described in /usr/src/gnu/usr.sbin/isdn/docs?
> 

right now 0.2 is in the pipeline (0.1 is now in the tree). This has been
ported to 2.1 and -current. All the activity is currenty restricted to
Germany, although Jordan is trying to get some of the supported cards.

Currently only cards from German manufacturers are supported (Creatix,
Teles, Dr. Neuhaus). AFAIK, nobody in the US has invested any effort.

> 2.  Has anyone done anything that will work with the
>     Motorola BitSurfer Pro?
> 

not that I know of.

> 3.  What are the advantages of a card over an external box?
>     Interrupt rate? Feature control?
> 

with a card you get the full B-channel speed (64 K/sec, 128 K/sec with
channel bundling). Unless you have a really fast serial port (> 115.2)
you can't get the full speed when you use channel bundling with an
external device. Note that the II stuff doesn't currently support
channel bundling, though, so this point is moot.

If you have a 16550A with FIFO then you'll probably get (this is a gut
estimate) 1 interrupt for every 16 characters or so. Most 486 or better
PCs can handle this easily.

For ISDN the interrupt load will vary drastically depending on whether
you have an active or a passive card. The active cards have a processor
on-board and don't interrupt the host CPU so much. The passive cards
(at least, the German ones) use a chip-set with a 32 byte deep FIFO,
so you get 1 interrupt for every 32 bytes, best case. Since the D-channel
packets tend to be small and cause an interrupt the load varies depending
on where the bytes come from.

I've used the ii0.2 stuff with a 486, ftp 7.5 kB/sec., and noticed no
degradation in performance.

An external box is probably easier to set up and use, since it looks like
a modem. So you can run slip/ppp with no changes to whatever comm. software
you like.

The II stuff uses raw HDLC. This is great if the guy you're talking to
supports it too, because the network is completely transparent
(the ii0.2 stuff does demand dialing, you just need to access the address
of the interface and, voila !, the connection is made). There's a sort
of AT command-set emulation, but I've never tried it out and can't say
whether it would work for slip/ppp. It definitely will NOT work for iijpp.

Feature control is almost non-existent in the II stuff.

---
Gary Jennejohn				(work) gjennejohn@frt.dec.com
					(home) Gary.Jennejohn@munich.netsurf.de
					(play) gj@freebsd.org





Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?9602231154.AA00568>