802.11 vs. 3G For Mobile Access

bobdole34 writes: "A new way to give us fast mobile net access spells further trouble for 3G.Imagine being able to surf the net at speeds faster than DSL from anywhere, at any time - you could watch a live video webcast while waiting for the bus, email photos to your friends while sitting in the park, or download the MP3 of the song that's playing in the pub before it finishes. I smelled vapour until I saw a demo of MeshNetworks at 802.11Planet in Philly."

Comments

You can't believe everything you see in demos

[pompeiisneaks]

Just because you saw a cool demo somewhere doesn't necessarily mean it STILL isn't vapourware... what if this company folds tomorrow?

sheesh, I saw tons of cool demos at linuxworlds in the past and nothing came of them!

Re:Where's the bandwidth going to come from

[yahwey]

The amount of bandwidth used depends on how many hops you are from the access point. You will use bandwith from every hop between you and the AP. So if most everyone has to go through say 5 nodes, the available bandwith will be saturated before you even get to the AP.

My question is what happens when everyone leaves downtown at night? Since there is no one to relay the signal, you'll only have coverage when you are close to an AP

Open the specs

[chris_mahan]

Will they open the specs?

Didn't think so. They are looking to boost stock value in an effort to capitalize on an eventual takeover.

If they did this for the good of mankind, they would have made the code open source.

What we need is a software-reconfigurable wireless lan card that could be programmed to communicate as needed...

Re: 802.11a = Too Much Power

[jandrese]

Color me dubious. The primary power drain is the _antenna_, at least on modern 802.11 cards. If you start cutting power to the antenna, you will start losing range. A better antenna design can offest this loss to a degree, but it is very expensive to manufacture extremely high quality antennas. What I can see is an adaption of the 802.11 that steps back your transmit power when you aren't using it to its fullest extent. There's no reason really to transmit at the full 11mbit when you're just sending HTTP Get requests and you're 5 feet from the basestation. If this is the case with the TI chip, then additional miniturization will not help your power consumption much. In fact the current jump would matter much less than the new sophisticated power control circutry added to the design.

802.x WAN will never get serious

[MobileDude]

It amazes me how people think of such things.

Anyone with experience in WAN coverage for Motient/Mobitex/RAM/ARDIS/Cingular/CDPD (whatever those groups are calling themselves this week) realizes that COVERAGE is KEY.

A convention center trade show painted with WiFi is not greater NYC. Just because you can surf the web with your laptop while on the second floor toilet in your house it does not mean you can scale outside your teeny D-Link range.

Let me guess - people will actually care to use a patchwork system that will have gaping holes every 1000ft or so. Yeah, sounds like a real world enterprise solution to me.

Get it straight - WiFi in the warehouse - WAN everywhere else (except the desolate Dakotas....).

Re:Plenty! Re:Where's the bandwidth going to come

[smallfries]

No it isn't.

Your first mistake is assuming that the quoted 11Mbs is the actual data-throughput. It isn't, because of the protocol overhead you lose 50% of the quoted bandwidth straight away. As the bandwidth is split into discrete amounts depending on your reception, you actually get 500kb,1Mb,1.5 Mb or 5.5Mb and unfortunately you really need to be in ideal line-of-sight for 5.5Mb.

Secondly, you assume that these huge numbers of people are all within range, of what? There are no good scalable peer-to-peer routing algorithms for mobile networks. The best ones in the literature all use a form of random flood filling of the network so on a city sized scale that's going to break badly. If you have base-stations everywhere then you need almost as much infrastructure as a 3G network. Add to that the really bad urban-canyoning and reflection problems and you end needing a basestation in nearly *every* starbucks ;)

We're actually in the process of rolling out a free 802.11 network over Bristol. Even though the project is funded the hassles of sorting out planning permission for aerials and putting in backbone means that after a year we've only just starting rolling with 3 aerials installed over a tourist area so that we can start doing demos.

Don't get me started on the other 802.11 limitations, it was designed for sitting browsing the web in the back-garden. It has latency and broadcast issues that make it unsuitable for streaming real-time video and audio, bad routing, IP hand-over issues and a pile of problems. People just look at the raw bandwidth figure and start creaming themselves over 55Mb bandwidth appearing everywhere but unfortunately the inherent problems with 802.11 mean that it isn't really as good as it sounds.