[wsfii-discuss] WiFiRe queries

Mon Oct 16 18:08:38 UTC 2006

Hello Fred, Ramon and others,

 Thanks a lot of taking the trouble to read the WiFiRe draft and also for
the encouraging comments.
Here are some responses to your queries. I have just now joined this list,
so I could not have replied earlier!

To summarize the queries:
> > - Why use only one channel?
> > - Has thought been given to lobbying for channel usage?
>* > -Is that draft something which is intended to be applicable to current
**deployed RF equipments 802.11b/g?
*>* > -Regardless of the response to the previous question, I assume that to
**make it successful, will require the support from the industry
**(manufacturers,
drivers...), is that right? There is already any **visibility on this?
*
Fred's understanding of the reason for a  single WiFi channel is correct.
Even though we are in the unlicensed spectrum, we do need to "reserve" some
spectrum for WiFiRe. We dont expect any significant interference since we
are using directional antennas and also expect to operate only in rural
areas. We have begun the process of lobbying with the govt.

 We havent yet given any thought to the 5 GHz band. We have chosen
802.11bprimarily because of the low cost of the chipsets and believe
that it should
not be difficult to override the wifi MAC while retaining the wifi PHY.
Preliminary discussions with potential industry partners indicate that this
should be possible.
The idea of using the neighboring 20MHz channel (to the de-licensed
channels) is good. We need to explore this.

Earlier, the TeNeT group at IIT Madras (www.tenet.res.in) had come up with a
rural connectivity technology (called corDECT), based on the DECT standard.
They have commercial deployments, which have been in use for a few years
now. WiFiRe builds on the learning gained from corDECT. The topology is the
same as that used for corDECT, so that we can leverage on the existing
infrastructure. So currently we are focusing on getting something going with
LoS and star topology.
 We havent started thinking about introducing nLoS features or using other
topologies yet.

If any of you folks interested in WiFiRe, are visiting Mumbai or Chennai, we
could meet for a discussion. I would have like to attend the airjaldi summit
but unfortunately it was too late by the time I came to know of it.

best wishes,
sridhar

---Fred's message is pasted below------

Hello Sridhar Iyer, Nagarjuna and alll,

I now have taken the time to reread the whole draft focussing on the
rationale of the proposed system, not the indepth MAC technological
details since that is not my field.

The most exciting part is the ambition to introduce WiMax (802.16)
concepts into the WiFi (802.11) MAC layer. This would create a "best
of both worlds" and could silence the ongoing waiting for and hyping
of WiMax while keeping the low-cost factor of WiFi. Making this
development open-source is very wise and will raise the level of
Indian expertise in wireless technology while creating opportunities
for the rural poor to use telecoms for their benefit. Although not
mentioned in the draft I hope that Near or No Line of Sight (NLoS)
features of WiMax can also be introduced into WiFi.

A great deal of the model is based on the existing cellular phone
infrastructure and assumes that the connectivity is supplied by the
exciting fiber optic infrastructure of these telecom operators and
hints at BSNL as the main supplier. Not only that but the model even
copies the cellular phone topology and tower infrastructure. The model
expands the range from 5km radius to 20km and increases the data rate
to broadband. For this to be allowed for WiFi the existing
restrictions of antenna height and EIRP have to be relaxed which would
mean some political lobbying and so the model introduces a form of
licensing in the just de-licensed 2.4-2.485 GHz band. From this point
I start to understand where the technological unusual choice for using
just 1 20MHz channel comes from. For the licensing to be political
acceptable other operators (than BSNL) will need to be given a chance
for a license as well thus occupying the available spectrum quickly.

Has thought been given to doing the same (lobbying for spectrum
relaxation) for the outdoor use of the 5 GHz (802.11a) band? Or
neighboring 20 MHz channels to the de-licensed 2.4-2.485 GHz band that
could be utilized by some low-cost WiFi chipsets?

The draft shows very good insights on costs of mast/tower heights
versus data rate and effects of those choices on the wireless topology
in the rural plains. In my experiments with antenna height in Goa
(flat land and a treeline even higher than 12m) I ran into LoS trouble
by the restriction for a maximum 5m antenna above rooftop. If some
NLoS features of WiMax could be introduced some of the cost aspects of
many medium high masts/poles needed for an other topology than the
proposed centralized star topology, like mesh or a grid, would also be
cost effective.

If some people from CEWIT can come to the AirJadi Summit in Dharamsala
to join us on a discussion workshop on these and other WiFiRe
subjects, we would be delighted to meet you.

For the participants at AirJaldi I would recommend reading the draft proposal:
http://www.cewit.org.in/docms/cewit06/WiFiRe-draft-lines.pdf
or if that is too long I put some of the background info on rural India on:
http://summit.airjaldi.com/wiki/index.php/Background.Rural.India
since the draft is CC I assume that doing that was correct.

Thank You,
Fred

On 10/13/06, Fred Pook <fredpook at gmail.com
<http://lists.okfn.org/mailman/listinfo/wsfii-discuss>> wrote:
>* Hi Ramon,
*>* Yes, this proposal would be feasible right now using atheros chips and
*>* something along the line of madwifi (http://madwifi.org/).  But i am
*>* sure anything could be rewritten in India for any hardware and be
*>* open-source without any commercial company involvement apart from the
*>* the low-cost hardware.
*>* Question for me still is: why use only 1 channel?
*>* thanks Fred
*>*
*>*
*>* On 10/13/06, Ramon Roca <ramon.roca at guifi.net
<http://lists.okfn.org/mailman/listinfo/wsfii-discuss>> wrote:
*>* >
*>* > First of all, sounds excellent to me any initiative oriented to solve
*>* > some of the limitations of the wifi specs which penalized it's usage on
*>* > medium distances.
*>* >
*>* > After a diagonal read, and I'm not the best in understanding some
*>* > technical questions about RF communications dropped in the draft but to
*>* > check my correct understanding of the overall and the impact that this
*>* > could have at the real world:
*>* >
*>* > -Is that draft something which is intended to be applicable to current
*>* > deployed RF equipments 802.11b/g?
*>* > -Regardless of the response to the previous question, I assume that to
*>* > make it successful, will require the support from the industry
*>* > (manufacturers, drivers...), is that right? There is already any
*>* > visibility on this?
*>* >
*>* > Sorry if I missed something obvious already described in the draft.
*>* >
*>* > Ramon.
*>* >
*>* >
*>* > En/na Fred Pook ha escrit:
*>* > > Hi Nagarjuna,
*>* > > I have done a quick scan of the WiFeRe specs:
*>* > > http://www.cewit.org.in/docms/cewit06/WiFiRe-draft-lines.pdf
*>* > > I find (among many great insides) a strange assumption to use only 1
*>* > > and the same 20MHz channel on all (3 or 6) sectored radio's on 1 base
*>* > > station mast. This would normally be done using several channels and
*>* > > different polarization. Using 1 channel creates the need for a new
*>* > > "shared" MAC layer and Time Division (what WiFeRe proposes).
*>* > > The choice for just 1 channel is based on avoiding other operators
*>* > > interference? (not clearly explained or motivated)
*>* > > There must be many great minds on the lists to explain if this a wise
*>* > > choice.
*>* > > Looking forward to the discussion,
*>* > > Thanks Fred
*>* > >
*>* > >
*>* > > On 10/13/06, Nagarjuna G. <nagarjun at gnowledge.org
<http://lists.okfn.org/mailman/listinfo/wsfii-discuss>> wrote:
*>* > >> Please see if any of you could review draft standard WiFiRe, ref in
*>* > >> the message.
*>* > >>
*>* > >> Nagarjuna
*>* > >>
*>* > >> ---------- Forwarded message ----------
*>* > >> From: Sridhar Iyer sri at it iitb ac in
*>* > >> Date: २००६ अक्तूबर १३ १३:३३
*>* > >> Subject: Re:
*>* > >> To: "Nagarjuna G." <nagarjun at gnowledge.org
<http://lists.okfn.org/mailman/listinfo/wsfii-discuss>>
*>* > >>
*>* > >>
*>* > >> Hi,
*>* > >>
*>* > >> I didnt know about this summit till I got your mail!
*>* > >> I am leaving for UK tonight for nearly 2 weeks, so I cannot attend.
*>* > >> I will see if any colleagues can make it.
*>* > >>
*>* > >> Are any of the trainers coming to Mumbai later? If so, we can try to
*>* > >> meet.
*>* > >>
*>* > >> Meanwhile, we have published a a draft standard WiFiRe (WiFi
*>* > >> Rural Extension) which defines a new MAC for an efficient rural
*>* > >> broadband access network using the WiFi Physical Layer. We have begun
*>* > >> some efforts to implement this and have made the standard open so that
*>* > >> others may also build equipment.
*>* > >>
*>* > >> This is published by Centre of Excellence in Wireless
Technology (CEWiT)
*>* > >> and available at www.cewit.org.in. We are currently in the review
*>* > >> process, so any comments you or others may have are welcome.
*>* > >>
*>* > >> best wishes,
*>* > >> sridhar
*>* > >>
*>* > >>
*>* > >> On Fri, 2006-10-13 at 12:09 +0530, Nagarjuna G. wrote:
*>* > >> > Dear Sridhar
*>* > >> >
*>* > >> > Are you or any of your colleagues planning to attend Airjaldi Summit?
*>* > >> > http://summit.airjaldi.com/
*>* > >> >
*>* > >> > regards
*>* > >> > Nagarjuna*
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.okfn.org/pipermail/wsfii-discuss/attachments/20061016/50a7d26d/attachment.html>