[TenTec] Why Ethernet?

[Duane Grotophorst]

--- Grant Youngman <nq5t at attbi.com> wrote:

> Of course, it isn't just Ethernet that's the issue. 
> Do you want to 
> Telnet to the radio? Do you want the radio to speak
> XML?  Do you 
> want the radio to serve up HTML so you can access a
> "soft panel" 
> from a thin client (e.g. your browser)?  And what
> about security if 
> you decide to offer yourself this access over an
> Internet connection, 
> especially if you're going to want to push the PTT
> from the remote 
> location? Lots of questions, that go well beyond
> basic issues of 
> Ethernet and TCP/IP.

All of these concerns are completely true. My vision
of the radio network is that it will be primarily a
'closed' environment. I myself am primarily interested
in having a network exclusively dedicated for my
shack. The wide area networking and station remoting
are icing on the cake that we can get to with minimal
additional work. But they really are secondary to
improving the 'in shack' experience.

I would design the radios themselves to be relatively
'dumb' devices (except for PC integrated/based radios,
- but then that is a different but related topic).
Their primary job would be to convert to/from RF/audio
(or digital data in the case of digital modes). So all
that they would be doing is streaming data In/Out that
contains 'audio' and control signals. They are really
not intended to be 'servers' in any sense of the word.
The 'smarts' of the user interface would be the
exclusive domain of the attached PC and/or the
knobs/button control panel. For over the Internet
scenario a gateway device would be required to provide
IP packet encryption and security.
 
> The idea of accessing and operating the radio over
> the internet 
> pretty much requires that you be able to serve up
> receive audio and 
> send transmit audio using voice packetization
> technology, and that 
> raises a lot issues, mostly related to delay, jitter
> and quality of 
> service, and their impact on audio quality.  Sadly, 
> it just won't work 
> reliably over today's Internet, period,  because (1)
> you can't control 
> what the other guy is the doing, and (2) the
> Internet (as opposed to 
> what you can do in your own home or corporate
> network) isn't 
> QoS/DiffServ enabled, which is absolutely essential
> for packetized 
> audio.
> 
> (This isn't true for "streaming" audio, but streamed
> audio for a 
> number of technical reasons (mostly jitter buffer
> size) won't work 
> well for typical two-way VOX-like conversations. 
> It's fine if you just 
> want to listen to your rig from grandma's house.

Also very true, however I think you can now achieve a
fair level of performance for voice rag chewing with a
broadband Internet connection, - about on par with
iLink or IRLP et al. I would have no illusions however
of ever pulling this off over a dial-up connection. On
the other hand if the public Internet infrastructure
improves to the point where there is near zero latency
connectivity, - then the Ethernet equipped gear in
your shack will be ready to take advantage of it, USB
will not.

The remoting capabilities that I see which would be
far more appealing now is the ability to have your own
dedicated fiber optic or microwave link to your remote
gear on the hill/mountain that you can see outside
your shack window. Or even having tower mounted
VHF/UHF/SHF gear fed with fiber to avoid having to buy
all that expensive 1-5/8 inch hard line.

> For in-station control, the USB idea is a pretty
> good one  -- and the 
> conversion from/to legacy RS-232 is relatively
> inexpensive and 
> simple.  You could do it today.

Again it goes beyond simple control. Even for control
only applications USB will not be any improvement over
RS232 for runs greater than a few 10's of feet, in
fact it is worse. While USB may seem to be the heir
apparent to RS232, it really is not the best choice
which will allow for the most versatility.

Duane
N9DG



__________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo.
http://search.yahoo.com