Research: Wi-Fi 'Inefficient Method of Communication' Networks, Particularly in Busy Areas, Less and Less Efficient According to new research being carried out by the University of Twente’s CTIT research institute, on behalf of Radiocommunications Agency Netherlands, the efficiency of Wi-Fi is degrading rapidly as more and more networks pop up -- particularly in urban markets. Most routers already deliver only a fraction of their marketed throughput, and the researchers found that in some busy locations -- the bandwidth available for actual data traffic can drop to less than 20 percent. "Wi-Fi has become an inefficient method of communication," argues Doctor Roel Schiphorst, who notes vendors need to work on improving future standard performance in busy locations. That's particularly important as more and more wireless carriers look to Wi-Fi to help ease cellular congestion.
|
 patcat88
join:2002-04-05
Jamaica, NY
kudos:1 | agree Yep. Wifi is unusable in most urban areas like NYC. 500 nodes 5 star nodes and none of them are connectable. | |
|  |  davoice
join:2000-08-12
Saxapahaw, NC | Re: agree If everyone would turn their transmit power down in those urban environments, down to say 30-40mW, things would be a lot better. The problem is you have lots of APs covering small areas but blasting 65, 80, 100, even 200mW of signal out way past the intended coverage range. Instead of learning to 'whisper' to reduce the roaring noise, everyone just keeps 'yelling' louder and louder.
}Davoice | |
| | | | | Re: agree I have mine set to 22mW and its fine. | |
|
Reviews:
 ·Fairpoint Commun..
| Wi-Fi Saves Battery Compared to Cellular In my Optimus V, if I leave my screen turned off but keep Wi-Fi on while still in airplane mode, my phone can last me about 14 days before the battery goes dead.
I'm unsure if I can compare this to cellular, because the far away the phone is from the cell tower, the more battery it gets used. Am I correct? | |
| |  Jim_in_VA
join:2004-07-11
Cobbs Creek, VA
kudos:4 | Re: Wi-Fi Saves Battery Compared to Cellular You are correct that more battery power is used to boost the transmission the further you are from a cell site
--
... need help? »evdo-tips.com/ | |
| | davoice
join:2000-08-12
Saxapahaw, NC | WiFi performance depends a lot on the AP. If the AP doesn't support power saving mode(s) (i.e. sleep/idle), which a lot of older ones don't, then you'll still be lucky to get a day or two out of it.
}Davoice | |
| | | davoice
join:2000-08-12
Saxapahaw, NC | Re: Wi-Fi Saves Battery Compared to Cellular Random geeky PS... If you remember the problems that Intel Centrino laptops had connecting and staying to connected to certain APs/routers when the Centrino first came out... well that was caused by the Centrinos trying to use power saving modes that didn't exist on the routers they were connected to. Intel's original drivers just blindly assumed that every router supported those modes (even though they were optional in the spec). So every time one of the Centrino radios would try to power save on an unsupporting router, it would disconnect and get stuck not being able to reconnect. Thankfully we've come a long way since 2005!
}Davoice | |
| | | | | Well, probably because I'm using an Atheros chipset with an ath5k module in Debian Sid. I'm using hostapd (wireless access point daemon). | |
| | | | davoice
join:2000-08-12
Saxapahaw, NC | Re: Wi-Fi Saves Battery Compared to Cellular My Optimus V does pretty well on my Cisco EA4500 too. I leave cellular turned on but run WiFi for data here at home b/c the Sprint signal is only 1 bar. I can usually go 3 days between recharges as long as the phone doesn't spend much time off my home WiFi.
I do notice however that it gives Sprint's authentication servers or VM's proxy servers a severe case of heart burn when a phone spends a lot of time off-cellular. Transitions to 3G from WiFi will fail after a couple minutes unless you reboot the phone. I'm running IHO Backside CM7 ROM on mine and that works better than the original stock ROM but it still fails often when the phone has been off-cellular/3G data for an extended period. I see 3G connect, it stays connected for a minute or two and then gets kicked off the network. That tells me that it's a network authentication issue.
(I'm on one of the grandfathered $25/mo plans with it. Plan to ride that out until Beta Wave C comes up with Republic Wireless.)
}Davoice | |
| | | | | Reviews:
·Fairpoint Commun..
| Re: Wi-Fi Saves Battery Compared to Cellular Well, I'm with T-Mobile recently since I have purchased a Samsung Exhibit II 4G for about $170 off of Amazon. I then put in CyanogenMod 9 (unofficial build from jocala at XDA Developers forum). So what is left with my Optimus V is I have it always in airplane mode all the time and it can do other functions like function as a SIP client using CSipSimple, use a remote control for XBMC, and do some other things.
I also have a SIP client in my Exhibit II 4G, but my primary SIP phone is my Yealink SIP-T22P, so my hardware phone and two Android devices are all connected to my Asterisk PBX. Because I don't use T-Mobile a lot, I'm only paying $3 for up to 24 hours which comes with unlimited voice minutes and texts but with 200 MB of 4G data, but T-Mobile could create a $3/day play that comes with 30 minutes, no text messages, and 500MB of data since I rarely make and receive phone calls when I'm not at home.
And I'd never thought about authentication servers when it comes to using cellular data.
--
Phone: Yealink SIP-T22P + CSipSimple in Optimus V
Phone System: Asterisk 10.1; Server: Debian Sid+Exp
I'm in heaven with VoIP except for 3G wireless. | |
| | | | | |  pfakPremium
join:2002-12-29
Vancouver, BC | Re: Wi-Fi Saves Battery Compared to Cellular said by GraysonPeddi:And I'd never thought about authentication servers when it comes to using cellular data. There are no authentication servers on the TCP/UDP part of the cellular data connection. If there is a transparent cache, it would have to have a Captive Portal to capture credentials. Otherwise, all that kind of stuff happens on the backend.
--
The more I C, the less I see. | |
|
Markie
join:2003-07-26
Kalispell, MT | Don't get it... This makes no sense! Wi-Fi isn't becoming less efficient, if anything the spectrum is being used more efficiently. We do need to see 5 GHz get more used. In most places 5GHz has free spectrum even as 2.4 GHZ is overloaded. Also, Wi-Fi should never have had channels 2,3,4,6,7,8,10,11,14. A 1,6,9,13 (no 13 in the US) channel plan, 20MHz spacing, like in the 5GHz band would have greatly reduced interference problems.Someone using, say, channel 3 completely wipes out 1 and 5 or 1 and 6, whereas co-channel users cause far lower impact... | |
| |  mmay149qPremium
join:2009-03-05
Dallas, TX
kudos:48 | Re: Don't get it... said by Markie:In most places 5GHz has free spectrum even as 2.4 GHZ is overloaded. Also, Wi-Fi should never have had channels 2,3,4,6,7,8,10,11,14. A 1,6,9,13 (no 13 in the US) channel plan, 20MHz spacing, like in the 5GHz band would have greatly reduced interference problems.
I totally agree, in fact I wonder why we never used unused spectrum from the beginning instead of the 2.4ghz spectrum considering WiFi used to severely interrupt older wireless telephones...
I think what we should be doing is educating the people on spectrum itself instead of marketing it the way we are, I've been to houses where people have 15 + devices connected to 1 WiFi router and then wonder why their internet runs like crap, or they can't transfer files unless they want to wait a day between devices. These people don't understand why their having these issues because they don't understand how spectrum and radio signals operate.
When you explain to people that wireless should only be used for dynamic devices, or that one static device that doesn't have a network cable accessible, and only use wired for transferring files (if you can help it) they start to understand it a lot more, instead of just thinking "I have wireless I can connect everything to it and everything will be fine!"
Edit: Forgot to add this, I think the other reason for these "inefficient" claims is because most places that serve free WiFi or etc get Enterprise level "Cisco" or other WiFi Access Points, that can only service maybe 50 - 200 people at once and are usually placed in a stadium or etc that 500 people are trying to access. What they really needed to do to fit the needs of people was to implement the equivalent of a cell tower for WiFi where it can accept up to 1,000 or so connections at once.
End Edit.
Matt
--
I am no longer an AT&T Employee. Check out my kudos! »/profile/1626573
Have U-verse questions? Please email uversecare@att.com and they will assist you!!  | |
| | | Markie
join:2003-07-26
Kalispell, MT | Re: Don't get it... 2.4 GHz was less crowded than 900 MHz, and 5 GHz doesn't have the range.
50-200 people on an AP? LOL no way. Perhaps in allowed associations, but not in use. All major vendors I know of recommend 30 users/AP except Ruckus. I've never used Ruckus but they're claiming far more.
It's not about how many connections are accepted, but about how well they can work. Ideally, in a stadium setup, each AP should cover maybe a couple hundred seats (30 users typically or so). With only three non-overlapping channels (four, but one of them isn't generally used in the US because it bumps up against a restricted band, so it has strict power and emission limits on it here) in the 2.4 GHz band, this means channels will get repeated too soon.
A more practical plan is to cover maybe one 2.4GHz AP per 500 seats, and then put two 5 GHz AP's in the same area. MUCH better use of the spectrum. The problem is...
MOST PHONES, ETC DON'T DO 5 GHz!
This is where 802.11ac will be great. 802.11ac spec mandates 5 GHz. I can't see deploying 802.11ac commercially. It just uses far too wide of channel for enterprise deployments. But it'll mean a lot more people who can use the 5 GHz 802.11n channels! | |
| | | | davoice
join:2000-08-12
Saxapahaw, NC | Re: Don't get it... Actually, with some of the newer carrier grade sectorized APs, you can get 1300 active users. It's kinda cheating though because I consider each sector to technically is its own AP but they don't since it uses a central controller and multiple radios.
}Davoice | |
| | | | | Markie
join:2003-07-26
Kalispell, MT | Re: Don't get it... Which ones? Link to products, and yes, that's cheating. 3x120 sectors is 3 AP's, not 1. Still, I'd love to see a product claiming 1300 active users on an AP. That just downright violates physics. 1300 connected maybe.
For example, on one random Cisco product I looked up, they say the AP *can* handle 2048 associated clients, but they *recommend* no more than 24: »www.cisco.com/en/US/products/hw/···3e.shtml
Just because it *can* handle 1300, doesn't mean it's recommended. Still, I'd love to read the product info! Which AP are you referring to? | |
| | | | | | davoice
join:2000-08-12
Saxapahaw, NC | Re: Don't get it... I'm surfing from bed at the moment. Will have to check my browsing history at work tomorrow for the make and model. This is actually something I was pulling together a bunch of research together on for a client. All I remember offhand is that it wasn't Cisco or Motorola, and that it had space in the ceiling mount enclosure for 12 or 13 radio sets (1 for each sector). The radios offloaded to a central routing/access processor/controller and used 3 GigE network connections for backhaul. Pretty neat implementation but I'm blanking on the name as the sleep meds start to kick in.
}Davoice | |
|
| | | mmay149qPremium
join:2009-03-05
Dallas, TX
kudos:48 | Oh yeah I totally agree, and my numbers weren't meant to be accurate they were just an exaggeration, you say that they only allow 30 at most, but I've worked in factories where at least 100 users at a time could be connected to one AP for the wireless bar code scanning equipment we had. So I know it's possible for more as I've seen and replaced the gear that Cisco makes for it. Either way we need gear that can handle more users, which is going to be hard to come by.
Matt
--
I am no longer an AT&T Employee. Check out my kudos! »/profile/1626573
Have U-verse questions? Please email uversecare@att.com and they will assist you!! | |
| | | | | Markie
join:2003-07-26
Kalispell, MT
2 edits | Re: Don't get it... They ALLOW more, it's just not recommended. For example Cisco recommends 24 users/AP. EnGenius (oddly, maker of much lower grade equipment but still decent) recommends 30/AP. Etc. Ruckus has higher recommendations. The limit on most professional equipment is adjustable, but CAN be set as high as 2048! It just slows to a crawl. Setting a limit of 30-50 users/AP, turning down the power, and adding more AP's (as well as 5GHz AP's to offload iPads and other 5GHz capable clients) has better performance. Barcode scanning isn't a high bandwidth application. Normal use, those 100 users would be slowed to a crawl.
P.S. I just looked, Ruckus says 100 data users or 20 VoIP users per AP radio. Meraki also says 100 users per AP, which I find surprising since there's no "magic" in the Meraki AP's I'm aware of (Ruckus uses smart antenna arrays to improve SNR) | |
| | | | | | Markie
join:2003-07-26
Kalispell, MT | Re: Don't get it... And to demonstrate that there's no magic secret sauce with the Meraki AP's, here's some testing showing the Meraki units being absolutely crushed in a high interference environment. Doing worse - much worse - than the consumer-grade Apple Airport Extreme. Ruckus however proves it really does have some secret sauce in it to handle these situations: »c541678.r78.cf2.rackcdn.com/revi···rt-2.pdf | |
|
| TheMGPremium
join:2007-09-04
Canada
kudos:1 | said by Markie:This makes no sense! Wi-Fi isn't becoming less efficient, if anything the spectrum is being used more efficiently.
That may be true, but as the number of WiFi enabled devices grows exponentially, the interference concerns are growing much more rapidly than the technology is improving.
I think part of the problem is way too many devices rely on WiFi these days, and the average consumer thinks WiFi is the ultimate solution to all their networking needs. It's not. WiFi should be used only for mobile devices (ie: phones, laptops, tablets). TVs, desktop computers, network media players, game consoles, printers, etc, really shouldn't be using WiFi.
Unlike what some people like to believe, the future is not completely wireless. Wired networking technologies will always have a place, even in household situations. | |
| | | Markie
join:2003-07-26
Kalispell, MT | Re: Don't get it... Oddly enough, I can say I totally disagree with you in home environments. In a single-family home, Wi-Fi works very well and interference concerns are minimal. Of course, I wire desktops when possible - why not? But I have no hesitation about the Wi-Fi only single family home.
For apartments, 5 GHz Wi-Fi is the promised land. More channels and shorter range! Shorter range is better in apartment buildings - less interference. 802.11ac will obviously add a whole freaking ton of interference, and I don't plan to deploy 802.11ac for business customers anytime soon, if ever. I need all those 20 and 40 MHz channels. But 802.11ac will stay mostly to a few neighboring apartments, dramatically improving Wi-Fi performance in those situations.
Wi-Fi is very efficient. Not quite (until 3x3) as efficient as LTE, or even the current UMTS releases, but pretty darn efficient, especially in 802.11n. 802.11n 3x3 MIMO (not widely used) is more efficient (getting up to 450mbps PHY in a 40MHz channel, whereas LTE Rel 8 2x2 pushes 100mbps PHY in a 10MHz channel).
Now, understand LTE gets much closer to seeing it's PHY layer speed at the application layer than 802.11n does so even 3x3 802.11n isn't as efficient as LTE, but seriously, with the very narrow re-use (Wi-Fi rarely covering even 100m - though I have AP's deployed providing outdoor coverage 200+ meters out from the AP). Due to this narrow cell spacing, the Wi-Fi spectrum is some of the most efficiently used spectrum in the United States. So how on earth anyone can claim Wi-Fi is inefficient is beyond me.
The problem is we're just demanding so much that 60MHz of usable spectrum (on the 1, 6, 11 channel plan) isn't enough. Easing operating restrictions to allow a 1, 5, 9, 13 channel plan on US equipment would be a potential step forward - but there's too many users out there on channels 6 and 11 (which shouldn't even exist... but that's another rant) to allow this to be of benefit in apartments and the like. But it'd be great for large campus deployments. Also, remember that another killer in apartments is people using adjacent channels. All it takes is a few uses on overlapping channels to totally ruin Wi-Fi for everyone. Adjacent channels are far worse than co-channel. Wi-Fi should ONLY have 1, 5, 9, 13. The other channels should have never been built. But 802.11b needed 22MHz - thus the 1, 6, 11, 14 scheme. Which was wasteful too - since that is 25MHz spacing for 1, 6, and 11 (and 22 MHz to 14).
5 GHz is the new frontier to fix this. There are no overlapping channels. That alone is a huge benefit. The range is much shorter. We just need to oh, actually start using it! | |
|
Sammer
join:2005-12-22
Canonsburg, PA | Inefficient control mechanisms All the article really tells us is that inefficient control mechanisms can use up more than 80% of Wi-Fi bandwidth. | |
| | | | Re: Inefficient control mechanisms Not to mention what the Wi-Fi has for backhual at that. If you feed it with say DSL or Cable vs. Fiber or Fios can make all the difference in the word on how may people can use it. | |
|
| | Some ideas The article refers to the amount of noise which causes collision collapse. Something wireless ISP's carefully plan and work around using GPS timed transmissions mixed with Time division multiplex.
If the wifi standard allowed the radios to lower their transmission power so that you followed the old ham radio rule 'dont transmit at a higher power than you need to' then it would work much better in crowded urban areas.
You will never be able to get more than 30 low speed users on an AP (or sector if it is an ap with sectored antenna) unless you have some sort of TDMA system implemented as part of the 802.11x standard, in both the AP and the client device.
I run a rural internet provider that uses a modified form of wifi to deliver high speed internet using 2.4ghz microwave radio over long distances. The radio manufacturers have added TDMA so i can fit up to 80 clients on a base station - but we still only have a total of 30mbit of throughput. Without TDMA and GPS timing we would only see about 3mbit because of collission collapse during peak hours.
Collision collapse is a result of CSMA. Early thinnet and thicknet ethernet networks suffered the same problem on 10mbit coax.
Wifi also now suffers from it.
802.11n was a good place to insert TDMA in place of CSMA but they didnt.
In 802.11n, they dropped the overhead from 50% (a 54 mbit 802.11g network really only puts through 24mbit) down to about 33% so a 64mbit 802.11n network can put through 42mbit
They could have also used the N standard to put in an 'efficient' mode. This would be where they drop the channel width from 20mhz to 5 or 10mhz, increasing the non-overlapping channels in the 2.4ghz range from 3 to 6 or 9. Of course you would get half the speed but for people only using wifi to extend a 10mbit DSL signal, a 5mhz (One quarter bandwidth) would still allow 11mbit through with a good signal on a single chain.
802.11n can make good use of spacial isolation - two chains (two antennas) tx/recieving on the same channel on the same frequency. One is horizontal and the other is vertical. You get twice the throughput without the extra bandwidth usage.
So now the idea of a 5mhz channel using one quarter of the radio bandwidth can put through double or 22mbit of usable data. The same as the old 802.11g on a 20mhz channel width.
Your laptop or client device is more sensitive when listening and transmitting when the channel width is only 5mhz wide instead of 20mhz, giving you 4x the range of the original 20mhz channel.
This extra range would have the potential to cause interference but can be offset by the fact there are more channels avaliable within the same 2.4ghz band and transmission power could be dynamically adjusted for when data is being transfered and to improve / reduce the signal level of the client device. | |
|
| |
|
|
