Hi all,
Today amny members turned up at HBCSE to put up an experimental mesh. Fred Pook walked us thru several aspects of deployment. We will be on the job on Sunday ironing out thingies. It was most interesting cause we fiddled with hardware which is what i love the most. Make it a point not to miss this most interesting meet.
Today amny members turned up at HBCSE to put up an experimental mesh. Fred Pook walked us thru several aspects of deployment.
Hey JTD
Is this Wireless Mesh same as the WiMAX word which is going around HOT these days?
Thanks and Regards Harshal Vaidya http://www.lakhpatipage.com Cheapest way to advertise on the internet.
__________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
Harshal,
Is this Wireless Mesh same as the WiMAX word which is going around HOT these days?
Umm Wireless Mesh would work on 802.11b/g (2.4Ghz) band, and usually has a very small range (which is why mesh needs to be done)
But, WiMAX on other side has a larger range (depending upon equipment used, it would give range upto 75 kms as well) This technology is very expensive at the moment to deploy, but has an excellent potential to outrun CDMA, GSM, WiFI everything (since it can beam speed in MBps for distant locations, this means you can see streaming videos over wireless)
You can read more of this stuff on http://www.wimax.com/education
Thanks & Regards, Mitul Limbani, Founder & CEO, Enterux Solutions, The Enterprise Linux Company (R), www.enterux.com
On Sunday 16 April 2006 02:13 pm, Mitul Limbani wrote:
Harshal,
Is this Wireless Mesh same as the WiMAX word which is going around HOT these days?
Umm Wireless Mesh would work on 802.11b/g (2.4Ghz) band, and usually has a very small range (which is why mesh needs to be done)
But, WiMAX on other side has a larger range (depending upon equipment used, it would give range upto 75 kms as well)
Not quite you are restricted by the earth's horizon whch is 41Km. apart from LOS restrictions. with high masts you may extend range by 10Km. In fact due to the higher frequency of operation it has poor penetration and will give u lesser coverage for the same power. However Wimax uses OFDM modulation which gives far better multipath performance and hence superior performance in dense areas and long distances.
This technology is very expensive at the moment to deploy, but has an excellent potential to outrun CDMA, GSM, WiFI everything (since it can beam speed in MBps for distant locations, this means you can see streaming videos over wireless)
You can read more of this stuff on http://www.wimax.com/education
Thanks & Regards, Mitul Limbani, Founder & CEO, Enterux Solutions, The Enterprise Linux Company (R), www.enterux.com
On Saturday 15 April 2006 16:30, jtd wrote:
Today amny members turned up at HBCSE to put up an experimental mesh. Fred Pook walked us thru several aspects of deployment. We will be on the job on Sunday ironing out thingies. It was most interesting cause we fiddled with hardware which is what i love the most. Make it a point not to miss this most interesting meet.
I hate you. I hate you. I hate you for evoking more of my interest :( Um...bad timing I guess but next week we have continuous practical exams and vivas :( I was keeping away from this highly interesting thread. May be you guys could demo this again in July? ( please ) Engineering students will be free only in June - July.
*curses Mumbai University's bad timing*
Dinesh,
Quoting Dinesh Joshi dinesh.a.joshi@gmail.com:
I hate you. I hate you. I hate you for evoking more of my interest :( Um...bad timing I guess but next week we have continuous practical exams and vivas :( I was keeping away from this highly interesting thread. May be you guys could demo this again in July? ( please ) Engineering students will be free only in June - July.
Well, this could be interesting, infact we can organise for something more interesting by then. How about having a handon with Asterisk running on top of Wireless Mesh ?
Any takers on this do respond. And yeah, pardon me for hijacking the thread :)
Thanks & Regards, Mitul Limbani, Founder & CEO, Enterux Solutions, The Enterprise Linux Company (TM), www.enterux.com
On Sunday 16 April 2006 05:52, Dinesh Joshi wrote:
On Saturday 15 April 2006 16:30, jtd wrote:
Today amny members turned up at HBCSE to put up an experimental mesh. Fred Pook walked us thru several aspects of deployment. We will be on the job on Sunday ironing out thingies. It was most interesting cause we fiddled with hardware which is what i love the most. Make it a point not to miss this most interesting meet.
I hate you. I hate you. I hate you for evoking more of my interest
:( Um...bad timing I guess but next week we have continuous practical
exams and vivas :( I was keeping away from this highly interesting thread. May be you guys could demo this again in July? ( please ) Engineering students will be free only in June - July.
*curses Mumbai University's bad timing*
-- Dinesh A. Joshi
Yeah, and I curse myself for health problems :(. I would certainly like this demo to be held again someday...
Thanks, Mrugesh
On Sunday 16 April 2006 05:52 am, Dinesh Joshi wrote:
I hate you. I hate you. I hate you for evoking more of my interest
Ok Ok u did not miss too much (if u look at end results). We managed to screw things up so that nothing was working when the meet started. However we did get everything almost up by 2100 (dhcp, all the APs talking to each other, clients connecting to their respective APs and obtaining ip). We broke up after that. we will be reconvening on staurday and sunday (Dr. Nagarjuna permitting) to continue where we left off.
:( Um...bad timing I guess but next week we have continuous : practical
Aha. Things were very hands on. Flashing the boxes, trying to get serial console on ones which did not have one. Jtag on one which would not boot. On Saturday Fred walked us thru the steps of site survey, antenna positioning, pros and cons of height etc. We used his GPS to plot obstruction positions and check out LOS conditions.
exams and vivas :( I was keeping away from this highly interesting thread. May be you guys could demo this again in July? ( please ) Engineering students will be free only in June - July.
YIKES. In rain?. Well why not we could checkout range in dry and wet weather. And what needs to be done to weather proof the enclodures.
*curses Mumbai University's bad timing*
The grass is green on the....
On Monday 17 April 2006 10:30, jtd wrote:
YIKES. In rain?. Well why not we could checkout range in dry and wet weather. And what needs to be done to weather proof the enclodures.
June is hardly a month for rains in Mumbai... :) BTW I read somewhere that you can dramatically increase the range of 802.11b/g devices by "modding" them a bit to add a home grown antenna. It doesnt cost much. May be we could explore this idea too. The bad thing about such an antenna is that it would be only directional.
On Monday 17 April 2006 11:05 pm, Dinesh Joshi wrote:
On Monday 17 April 2006 10:30, jtd wrote:
YIKES. In rain?. Well why not we could checkout range in dry and wet weather. And what needs to be done to weather proof the enclodures.
June is hardly a month for rains in Mumbai... :) BTW I read somewhere that you can dramatically increase the range of 802.11b/g devices by "modding" them a bit to add a home grown antenna. It doesnt cost much. May be we could explore this idea too.
Fred had two +12db omnis which were mounted on pipes. You need to mod netgear's AP not linksys to fit an external cable and antenna. The reflashed APs allow u to boost power to max (100mw).
The bad thing about such an antenna is that it would be only directional.
No. Only a directional antenna is directional in the horizontal plane. The omnis are directional in the vertical plane (pancake radiation - very little vertical radiation). The ones you are talking about are pringle's (tin can). They can give you massive distances of several kilometers. If u are very careful in the construction upto the horizon (40km). In a mesh you are interested in spreading, so omni it is. You can use a splitter and one omni and one directional to talk to another cluster. U can also use additional PA to pump the power. Fred has given me some PA chips. I will have them ready in 12 days for more fun.
jtd wrote:
Fred had two +12db omnis which were mounted on pipes. You need to mod netgear's AP not linksys to fit an external cable and antenna. The reflashed APs allow u to boost power to max (100mw).
Could you give a brief idea of what this mesh was all about and how different would it be from APs used in repeater/relaying mode to increase range? Sorry if I have missed out on something. What about bandwidth? Would it get divided like a serial/cascading load?
Regards,
Rony.
___________________________________________________________ Switch an email account to Yahoo! Mail, you could win FIFA World Cup tickets. http://uk.mail.yahoo.com
On Monday 17 April 2006 11:37 pm, Rony Bill wrote:
jtd wrote:
Fred had two +12db omnis which were mounted on pipes. You need to mod netgear's AP not linksys to fit an external cable and antenna. The reflashed APs allow u to boost power to max (100mw).
Could you give a brief idea of what this mesh was all about and how different would it be from APs used in repeater/relaying mode to increase range?
Most APs can work as repeaters or APs. So you cannot have a repeater acting as AP. Also a repeater will talk to one more repeater and AP. In a mesh network each ap can act as ap and repeater and will talk to whoever connects. It also has routing protocol OSLR which allows it to route packets correctly. This is adhoc mode.
Sorry if I have missed out on something. What about bandwidth? Would it get divided like a serial/cascading load?
Throughput deteriorates as the mesh size increases. U have increased latency and reduced per user bandwidth. These are things which are being addressed. Latency is the issue for voip over 802.11.
jtd wrote:
Throughput deteriorates as the mesh size increases. U have increased latency and reduced per user bandwidth. These are things which are being addressed. Latency is the issue for voip over 802.11.
This is a suggestion. If rural or long distance coverage with good latency and bandwidth is the final goal and cabling is too expensive, then a combination of wireless radio (APs) as well as wireless optical technologies could be tried out. The long link in lieu of the cable can be established using laser tx/rx units mounted high on relaying towers so that maximum LOS distance can be achieved. The last mile connectivity can be provided by the APs.
Optical trancievers will provide almost similar high bandwidth as the optic fibre cables and can be mounted on independent towers that are solar powered during the day and rechargable batt. during the night. A thin long laser beam should be adequete enough to reach a few kms. In order to overcome the wind sway of the tower tops, the optical sensors can be mounted at the focus of parabolic reflectors that are sized as per the maximum deviation of the beam due to the sway. A more sophisticated approach could be to use (if available and not expensive) the military type sensors used in laser guided weapons systems to be able to track moving beams. The entire assembly can be mounted in a sealed all weather box and a water pump positioned to automatically spray water at an off peak hour can be used to clean the optical system everyday, thus eliminating the need for a human touch every day for every tower. Water can be made available once a month from a tanker or from an underground tube well.
The protocol used could be such that a minor interruption caused by bird flights does not disconnect the link, it only pauses till signal is back.
At the final destination, a cluster of towers could optically connect with the final tower and this cluster will mount the APs to give full area coverage. In all this process, no cables are used.
I don't know if this idea may work but an experiment can be carried out in laser based optical connectivity between two tall buildings in Mumbai.
Regards,
Rony. Send instant messages to your online friends http://uk.messenger.yahoo.com
On Tuesday 18 April 2006 09:20 pm, Rony Bill wrote:
jtd wrote:
Throughput deteriorates as the mesh size increases. U have increased latency and reduced per user bandwidth. These are things which are being addressed. Latency is the issue for voip over 802.11.
Optical trancievers will provide almost similar high bandwidth as the optic fibre cables and can be mounted on independent towers that are solar powered during the day and rechargable batt. during the night. A thin long laser beam should be adequete enough to reach a few kms. In order to overcome the wind sway of the tower tops,
This is the problem. Not just wind but ground vibirations from passing vehicles, causes the link to loose connection all the time. Beside scattering due to rain and dust storms.
the optical sensors can be mounted at the focus of parabolic reflectors that are sized as per the maximum deviation of the beam due to the sway.
You need a very expensive servo head at both ends.
A more sophisticated approach could be to use (if available and not expensive) the military type sensors used in laser guided weapons systems to be able to track moving beams. The entire assembly can be mounted in a sealed all weather box and a water pump positioned to automatically spray water at an off peak hour can be used to clean the optical system everyday, thus eliminating the need for a human touch every day for every tower. Water can be made available once a month from a tanker or from an underground tube well.
Bore well water is full of calcium and residues will make the lens opaque faster than dust.
I don't know if this idea may work but an experiment can be carried out in laser based optical connectivity between two tall buildings in Mumbai.
Not cost effective at all. Wireless PtP links are far better. But from an academic point it would be very interesting.
-
Dinesh Joshi -
Harshal Vaidya -
jtd -
Mitul Limbani -
Mrugesh Karnik -
Rony Bill