[Cu-wireless] Re: WET11
David Young
dyoung at pobox.com
Fri Jan 24 00:39:18 CST 2003
If I understood your diagram right, each of the two routers has an
interface on 10.1.2.0/24. The first interface is ethernet; plugged into
it is the WET11. The interface on the second router is wireless. Correct?
Does the WET11 have an IP number in this whole scheme? The second router
does not accidentally addressing the WET11 by its IP or MAC, does it?
I would be interested to see a trace of what is going on at layers 2 and
3 on both the wired and wireless side. Will you produce a trace to share?
We may adopt IPv6 sooner rather than later. I wonder if the WET11 will
block IPv6?
Dave
On Thu, Jan 23, 2003 at 10:06:45PM -0800, Ken Caruso wrote:
>
>
> Hello,
>
> I would like to suggest that put the wet11 through the wringer
> before putting it into production. Seattlewireless.net has a very similar
> model/plan to what CU-Wireless is doing and we have run into some issues
> with the WET11.
>
> The big one being actually routing *through* the device. Here is a
> scenario:
>
> 10.1.1.0/24--|router|
> |
> |--10.1.2.0/24--|WET11|
> |
> |
> |
> |linux box w/ wireless card|--10.1.3.0/24
> |
> |
> 10.1.4.0/24
>
> In this case some coming from the 10.1.2.0/24 network (behind the wet11)
> can access all the networks that are connected to the linux box. I will
> make a pretty general statement, The WET11 seems to have no issues passing
> traffic to *any* network as long as it is on the wireless side. However
> if you put a routing hop *behind* the WET11, such as the 10.1.1.0/24
> network in the picture above, we have found it to be unreachable. That is
> a packet comes in on the wireless card in the wet11 but never makes it out
> the ethernet side. So basically it appears as though the WET11 will not
> pass traffic to the ethernet side unless it has a destination address of
> the local network physically connected to the ethernet port. I have not
> tried this with the latest firmware as I was pretty fed up with the device
> after that point. So it maybe it has been fixed. But I would suggest
> really testing out all applications other than its intended one which is
> to make clients that cannot accept wireless cards, connect to an Access
> Point.
>
> my 2 cents
>
>
>
>
> Ken Caruso
> kenc at seattlewireless.net
> http://ken.ipl31.net
>
> "when in doubt tell the truth" -Mark Twain
>
> On Thu, 23 Jan 2003, David Young wrote:
>
> >
> > What nodes? =)
> >
> > Seriously, we have more install candidates near Urbana High School than
> > we have nodes. And then there is the 400 block of Green Street, where
> > I think we could install at three more homes, easy.
> >
> > Incidentally, I am still looking for an inexpensive local distribution,
> > and I am not finding it. Any ideas?
> >
> > We have cable and connectors for a few installs. We need radios and
> > antennas.
> >
> > I want to survey on Ivan's roof just as soon as the snow has melted from
> > it. I forgot to check it this AM. Ivan's is an iffy install because we
> > will have to use a lot of cable or an outdoor router. The outdoor router
> > is more than we can afford.
> >
> > The LinkSys WET11 is a tiny ethernet<->wireless bridge with decent
> > output power and mediocre receive sensitivity (but mediocre beats
> > crappy any day). Detaching its antenna exposes an RP-SMA connector.
> > If it will not already tap ethernet for DC power (not the laboratory
> > namesake, but the electricity), then we can build it a tap using plans at
> > nycwireless.net. Then we can buy/build the injector. I think that if we
> > "pull a ham" and put it in Tupperware with holes punched for ethernet
> > and radio connector, then we've avoided excessive cable losses on the
> > cheap. Maybe....
> >
> > We still need to attach a PC to provide the routing/forwarding function,
> > and we should to verify certain properties of the WET11, such as its
> >
> > 1 power requirements (50 feet Cat5 turns a lot of electricity
> > to heat, but a 48VDC injector could cook the thing at 75)
> > 2 its radio properties (I think that it is superior to our WMP11's,
> > with 100mW and -83dB RX sensitivity, but maybe I am wrong)
> > 3 its configurability for channel and SSID (I prefer that it is
> > Web- or SNMP-configurable. I have read that it will start a Web
> > configurator if you give it an IP by DHCP, but seeing is believing.)
> > 4 its ability to handle an IBSS merge (this should not be a problem
> > if it has a Prism MAC with recent firmware, but you never know)
> > 5 temperature range
> > 6 software reliability (does it conk out after N hours or Q packets?)
> >
> > If somebody on the list already owns one of these, maybe you will put
> > it through its paces? If not, I will front the lesser of $50 or half the
> > purchase price for a WET11 if somebody will evaluate its suitability by
> > those criteria.
> >
> > Dave
> >
> > On Thu, Jan 23, 2003 at 10:34:43PM -0600, Perry Melange wrote:
> > > Hey, I have just gotten 19 pieces of 72 pin SIMM memory for us to use on
> > > the nodes. Do we want to get together some time soon and set up a few
> > > computers?
> > >
> > > Perry
> > >
> > > --
> > > Perry Melange "Make things as simple as possible, but no simpler."
> > > pmelange at null.cc.uic.edu - Albert Einstein
> > > http://null.cc.uic.edu/~pmelange/resume
> > >
> > > _______________________________________________
> > > Cu-wireless mailing list
> > > Cu-wireless at lists.groogroo.com
> > > http://lists.cu.groogroo.com/cgi-bin/listinfo/cu-wireless
> >
> > --
> > David Young OJC Technologies
> > dyoung at ojctech.com Engineering from the Right Brain
> > Urbana, IL * (217) 278-3933
> >
> > _______________________________________________
> > Cu-wireless mailing list
> > Cu-wireless at lists.groogroo.com
> > http://lists.cu.groogroo.com/cgi-bin/listinfo/cu-wireless
> >
> >
>
> _______________________________________________
> Cu-wireless mailing list
> Cu-wireless at lists.groogroo.com
> http://lists.cu.groogroo.com/cgi-bin/listinfo/cu-wireless
--
David Young OJC Technologies
dyoung at ojctech.com Engineering from the Right Brain
Urbana, IL * (217) 278-3933
More information about the CU-Wireless
mailing list