From: Paul Gardner-S. <pau...@gm...> - 2010-07-24 20:25:36
|
Hello Alexander, On Sun, Jul 25, 2010 at 5:01 AM, Alexander Chemeris <ale...@gm...> wrote: > Paul, > > What I worry the most reading this is power consumption. Mesh networks > tend to consume more power then infrastructure based. So while it's a > great thing for grid-powered Mesh Potato, it may be too power hungry > for mobile phones. We have put some thought into this. 2.4GHz WiFi is really not that great on this front, especially since it is not always possible to disable beacons and other bandwidth and power wasting features. Also, WiFi tends to be fixed power, rather than ramping the power down for nearby nodes. This is unfortunate, since in principle a mesh theoretically can have lower overall power consumption to carry a signal a given distance, since the 1/(n^2) rule dictates n^2 power for a given distance. But there are some things we can do. For example, as all the phones have GPS* you could easily synchronise them and use TX and RX slots allowing the WiFi radio to run at a low duty, say 10% or so when not carrying calls. This would extend the battery life to something respectable, although still not quite as good as GSM. I have also not yet put any effort into optimising the radio strategy of the BATMAN meshing software. This is another reason why I like ISM915 -- larger range implies less frequent beacons from the meshing software to maintain toplogy due to the increased range:velocity ratio of devices in the network. * Of course, we want to be able to work with zero infrastructure, so GPS would not be our only method here. It would in fact be possible to get nodes within a local area of a mesh to self-synchronise sufficiently for this purpose. > Have you estimated how power consumption will > increase when you turn on mesh networking feature? We don't have clear figures on the exact consumption yet, partly because we need to correctly adjust for the GSM radio sucking power, especially when out of range of towers. This is something that is high on my list, but at present my challenge is best summed up as a lack of time. As a general guide, though, we are seeing somewhat, but not excessively reduced, battery life on a HTC Dream with our stack running. > PS People say $1M is not too much if you have a great idea :) I agree, and that is why I am aiming to do this. We have already had strong interest from a number of organisations, and expect more. >From a national government perspective it seems compelling to me: We have already demonstrated the key technologies, from here is is incremental: better air interface, better user interface, etc, but the benefit of the resulting capability is revolutionary. If you or anyone can think of any organisations that we should engage with, do let us know. I know it sounds pithy and hyperbolic, but mesh mobile telephony that can truly work any where, any time, and can allows a citizen to setup a functioning interoperable mobile network in minutes using hundreds not millions of dollars, will result in change. Paul. > On Sat, Jul 24, 2010 at 23:18, Paul Gardner-Stephen > <pau...@gm...> wrote: >> Hello David, >> >> On Sun, Jul 25, 2010 at 2:05 AM, David A. Burgess <dbu...@jc...> wrote: >>> Paul - >>> >>> I think the idea of building a handset for an ISM band is interesting, but I >>> don't think that 915 is a good choice. First, the bandwidth may be too >>> narrow to support the frequency duplexing that is common in most cellular >>> protocols. Second, such systems would be useful only in ITU region 2 and a >>> few other oddball countries out there, like Australia, who don't strictly >>> follow the standard ITU bandplans. >> >> I hear what you are saying, but I still strongly believe that there is >> value in doing so. >> Partly this is because the main use of the facility would be in >> situations where most or all of the clutter would not be present, >> namely: >> >> 1. In remote/undeveloped regions, helping the poorest two billion >> actually get affordable access to telecommunications. >> 2. In disaster/emergency situations where existing infrastructure is broken. >> >> These are the two main target applications of servalproject.org, and >> ISM915 would be ideal, again because it can use the existing radio >> gear in the phone. >> >> I acknowledge that ISM915 is not universally available, but it is >> reasonably available. Something that covers the USA, Australia and >> South Africa has clear potential for demonstrating the concept, which >> can then be leveraged to find better bands or bands for other regions. >> Much of the effort will be readily transferrable to any new band, >> that is the beauty of SDR. >> >> Also, as the Chinese saying goes, the best time to plant a tree is 20 >> years ago, the next best time is today. I want to get the tree >> planted as soon as possible, so that the poorest in the world do not >> have to wait any longer than is necessary for the fruit. >> >>> A better band might be 2.4 GHz. There's plenty of bandwidth and it's much >>> more standardized across the globe. The down side is that you lose about >>> 60% of your range by moving to a higher frequency. You might make some of >>> that up with higher antenna gain at the basestation. >> >> 2.4GHz has its place, and we are already using it via WiFi ad-hoc >> mode, which is a good least-effort entry point. We already have >> phones running our software stack in that mode that can call one >> another with out any supporting infrastructure. We demonstrated this >> in the outback a few weeks ago: >> >> http://www.youtube.com/watch?v=K819ggBvkU0 >> http://www.youtube.com/watch?v=ZG0ph6zXa7s >> >> The without infrastructure is a key point of our technology and our >> mantra: Telecommunications any where, any time. Thus we cannot assume >> a basestation will be present, although the mesh will automatically >> make use of any node in a good position, e.g., on a hill top or >> hanging from a weather balloon. Also 2.4GHz has dreadful multi-path >> properties in heavily vegetated areas compared with 915. >> >> So in short, I agree, 2.4GHz is very useful, but I do not think it is >> the only band required to satisfy all use-cases, just as 915 is not >> the band for all occassions. >> >>> Your other big problem, in any ISM band, is that you will suffer horrible >>> interference from 802.11 networks, other unlicensed IP radios, CCTV >>> networks, remote-controlled airplanes and all kinds of other random junk, >>> especially if you try to operate over any significant distances. >> >> This is true, but again varies with the use case. The main value I >> see in adding 915 as a string in our bow is that it does allow longer >> range in remote/sparsely populated regions where the clutter floor >> will typically be low enough for a 1W signal to propagate a useful >> distance. The 1W allowance in most 915 jurisdictions is also helpful >> in increasing the maximum range. >> >>> Cellular >>> networks run on licensed spectrum for good reason: you need to be able to >>> control you uplink band over very large areas. It's not all about greed and >>> social control. Really. >> >> I do understand this. The key is that they offer a carrier grade >> service. Serval is all about providing the tools to create the most >> effective best-effort service in any situation. Therefore, 915 can >> only be an asset to us. >> >> Also, with mesh telephony you don't need to control the spectrum over >> a large geographic area in the same way that a telco does. Local >> interference is only a local problem, and if the software can bridge >> between radios (which it can) then the mesh will form the most robust >> network possible within the spectral resources available. >> >>> All that said, if you really want a handset for a non-standard band, it's >>> probably possible without "rolling your own". You just need to know a >>> handset maker and China and be prepared to order at least 10,000 of them. >> >> This is an option I would seriously consider to prove our point. >> I just need to get the million dollars to do it. >> >> Perhaps I am crazy for having this vision, but I don't think it >> matters. It is possible. It will help many people. As ambitious >> and imperfect as it might be, we are going to try, and we welcome the >> support and involvement of all that see value in our vision. >> >> Paul. >> >>> -- David >>> >>> On Jul 23, 2010, at 5:28 PM, Paul Gardner-Stephen wrote: >>> >>>> Hello, >>>> >>>> I would also like to pitch in with my support for a software mobile phone. >>>> >>>> In addition to the obvious GSM applications, I see a further benefit >>>> in that it allows use of the ISM915 band that lies between the up and >>>> down bands for mesh telephony. >>>> At servalproject.org we already have a P2P mesh mobile telephony >>>> architecture which we have proved, and could be used as an >>>> interconnect for networks of OpenBTSs. >>>> In fact, as soon as I can find some funding to do so, we will begin >>>> working on doing just that. However, the ISM915 side of things does >>>> require the same kind of crazy baseband processor coding as would >>>> implementing the GSM protocols themselves, but the value would be in >>>> the extended mesh range that would result. >>>> This could be used for stand-alone mesh mobile networks, but also to >>>> extend the range of the OpenBTS by using GSM as far as possible, with >>>> the mesh forming an annulus of extended best-effort coverage beyond >>>> that. >>>> >>>> If anyone is interested in doing their PhD in this space, or otherwise >>>> contributing to such an effort, I would be delighted to hear from you. >>>> >>>> On a related note, what does the OpenMoko use for the baseband radio? >>>> It it open or closed source? >>>> >>>> Paul. >>>> >>>> >>>> ------------------------------------------------------------------------------ >>>> This SF.net email is sponsored by Sprint >>>> What will you do first with EVO, the first 4G phone? >>>> Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first >>>> _______________________________________________ >>>> Openbts-discuss mailing list >>>> Ope...@li... >>>> https://lists.sourceforge.net/lists/listinfo/openbts-discuss >>> >>> >>> David A. Burgess >>> Kestrel Signal Processing, Inc. >>> >>> >>> >>> >>> >> >> ------------------------------------------------------------------------------ >> This SF.net email is sponsored by Sprint >> What will you do first with EVO, the first 4G phone? >> Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first >> _______________________________________________ >> Openbts-discuss mailing list >> Ope...@li... >> https://lists.sourceforge.net/lists/listinfo/openbts-discuss >> > > > > -- > Regards, > Alexander Chemeris. > http://www.fairwaves.ru > |