[MLB-WIRELESS] [TECH] Dipole antennas, and melbwireless structure

[Ben Anderson]

> > > > > And these routes can be statically specified.
> > > >
> > > > well, semi-statically...  static stuff sucks ;)
> > >
> > > Yeah, I'm trying to minimise traffic here, and if the
> > network can safely
> > > assume a particular route is up, all the better.
> > > Sure, though making a black-hole by having a flakey shortcut
> > sounds like a
> > really great way of making a network that just flakes all the time.
>
> I'm looking at a realistic case where someone can provide a shortcut, but
it
> may only be available for x hours a day (e.g. evenings).  Can anyone say
> "shared rent"?  "Arguments over power bills"? :-)

Yes, there will need to be a way to update the shortcut list.  However
advertising a shortcut to the network at large by any node that may come
online sounds like a really good way of DoSing this mechinism.  And as more
and more nodes come up with shortcuts, this list becomes prohibitivly large
in the same way it did before we had physical location based addressing (aka
mobilemesh).

> >
> > > > The long/lat stuff is useful in preventing the *whole*
> > > > network from being
> > > > discovered by broadcast...
> > >
> > > Yes, it will be more useful for knowing where to route
> > traffic destined
> > for
> > > distant nodes.  Local nodes have to be discovered.
> >
> > It's a broadcast zone.  any transmission implicitly implies
> > this discovery.
> > The discovery can be incidental, rather than explicit.
>
> Agreed, the discovery of something "missing" is always going to be slower,
> of course.  Why didn't the other end respond?  Because it was turned off?
> Because there was a collision?  Because it crashed?  Because it was at
100%
> CPU and couldn't respond at the time?   In the meantime packets are
getting
> delayed until it is known (or assumed by the expiration of a timer after a
> few retries) that the node has indeed died.  If the node can tell the
world
> (as in the case of a normal shutdown), all the better. :-)

Better, but we shouldn't rely on it.  And that's something that can be added
'later' easily in the design (ie via any means available, get the
information through to the source node -- he's doing source routing
remember....)

> > Yup, I was just pointing out it's unlikely to be trivial, and
> > thus, less
> > likely to get done.  Something worth considering in look at
> > how this will
> > ultimately scale.
>
> I feel Win32 is ultimately necessary.  Not everyone is going to run *NIX
on
> their node for whatever reason (ok, you may not like that, but some people
> have very valid reasons to use Windows), for this to become acceptable on
a
> broad scale, unless it becomes an appliance thing (unpack node, chuck on
> roof with omni and turn on sort of thing)...

I'm aware of that, know that, and indeed currently I'm using the ultimate of
all evil email clients to have this conversation.  Outlook express.
Just writing network layer drivers seem 'on the impossible' side of possible
to me at the moment.  Which is why I conceived the pc104 style embedded
solution.... black box with ethernet port and antenna, coupla hundred bucks.
Plug and play.
Doesn't cater for ipaq users (and I'd like to be one of these) but then,
there have to be some limitations...  run linux on the ipaq and it'll
work... writing CE network stack doesn't sound fun, and i'm not sure if it's
even possible since quite often CE is run from rom.

> > > What are you in this for?  Greed doesn't catch my eye. ;-)
> >
> > I've dreamt of a UPN for a very long time.  Thinking about
> > this stuff is
> > fun, making something work is a major buzz.  I also enjoy
> > having my ego
> > stroked by people telling me how much of a gun I am.
> > Being rich affords one a lot of scope to re-implement the
> > idea a lot better,
> > and faster.  Cutting through red tape is much easier with
> > money.  Money has
> > it's uses, but it's not something I'm greedy for.  (if I
> > were, why the hell
> > would I be typing such long emails? ;)
>
> Oh,I like the UPN idea, don't get me wrong on that score. :)

I know, I know :)

> > > <loads up a few nukes...> ;)
> >
> > heh, period of very hot nuclear fire, period of very cold
> > nuclear winter.
> >
> > I really should get some sleep...  And not type email when
> > I'm drunk...
> > That rich comment was very inflamatory...  i intended it with
> > so much toung
> > in cheek... Ahh well...  Can't be perfect at everything, all the time.
>
> Hehe, the problem with me is I've worked hard to get the priviledge of
> spectrum access that only a minority enjoy (but anyone can do, with some
> effort), and nowadays, commercial interests are convincing governments
with
> lots of $$$ to take that bandwidth away for their purposes.  Sure, public
> networks will cause some interference, but as long as they stay non
profit,
> we will all (mostly) get along.  As soon as it becomes big bucks, that
could
> change and we might all be kissing 2.4 GHz goodbye, unless we have $$$ to
> subscribe to ABC Wireless Internet, or live with flea  power, no external
> antennas and whatever crap the big boys spew into the band. :-/

I've wanted to do the HAM thing for years.  But in reality, the flea-power
type stuff has been able to satisfy my needs for bandwith till recently.  I
just need more time...  and some ham buddies probably wouldn't hurt either
*grin*

> The one saving grace is it's an ISM band, so interference from industrial
> (good old Microwave ovens!), scientific and medical equipment can be
> expected (and these applications are protected, I think as well), but that
> may not perturb big corporations with deep pockets and healthy research
> budgets...

Yup, and oxygen absorption band, and any other band where there's a really
crappy reason to want the bandwidth.
One thing that I think we should progress towards is a generic digital radio
service...  Everythings low-interferance, spread spectrum type technology.
Current services are moved across to the digital transmission, compressed.
Extra bandwidth is used to help relieve some of the bandwidth pressures that
currently exist.  In australia, we have low population/area, radio bandwith
shouldn't be a super-huge problem to have lots/capita.  I think it's just
not optimally managed.  (again, long term proposition.  I realise it's
difficult to just turn off AM & FM bands, analog TV stations, etc.  The
progression should be made...  And with TV, it seems that it's heading in
the right direction.  Free to air TV using digital exclusivly would free up
quite a lot of bandwidth (!)

> Food for thought - the ham band in this part of the spectrum used to be
150
> MHz wide... Now 98 MHz of that has been turned over to MDS TV broadcasting
> (2302 - 2400 MHz).

Yes, and I find this dissapointing.  Though at the same time, it's an
interesting moral issue to try and justify...


> UPNs, yes, but let's keep this a community project.

I don't think I ever proposed, at least I never intended to propose, that we
'sell out' to make this work.  Community == good.

Ben.


--
To unsubscribe, send mail to minordomo at wireless.org.au with a subject of 'unsubscribe melbwireless'  
Archive at: http://www.wireless.org.au/cgi-bin/minorweb.pl?A=LIST&L=melbwireless
IRC at: au.austnet.org #melb-wireless