Slashdot Mirror
7Gbps Wi-Fi Networking Kit Could Launch In 2010
Mark.JUK writes "Wireless Local Area Networking (WLAN 802.11) adapters capable of speeds 'up to' 7Gigabits per second could be in stores by the end of this year. The Wireless Gigabit Alliance (WiGig), which seeks to advance the worldwide adoption and use of 60GHz wireless networking technology, has published a unified specification for its approach and opened an Adopter Program. The move means that WiGig members can now begin developing a Wi-Fi kit that uses the unlicensed 60GHz spectrum."
"Could".
Will this "new, magical and unicorn-like" WiFi travel further? Far enough for municipal WiFi to effectively cover its citizens? If so then the increased coverage is more important than the speed improvement (even though the speed bump is might impressive).
What's the real bw available? 2 Gbps?
With 802.11n we get max 90Mbps from the carrier's 300; that's only 30% eficiency. I hope it's better this time.
It doesn't matter if you're throttled. I barely use bandwidth, and I'm still throttled all to hell.
Life takes interesting turns, but the most interest is when you're off the beaten path.
Faster networking speeds in the home and office (and coffee shops, I guess) is always a good thing, and we should hope that technologies that bring this about continue to progress.
But the real problem for many mobile users is networking speeds outside the office. At customer sites, in transit, and during leisure time activities, having fast, reliable network access is still a dream. You can expect slow and unreliable cellular service most places, but it's like stepping back to the bad old days of 56kbps to use the cellular data network.
More WiMax and other truly ubiquitous wireless networks are what is really needed at this time. That means both support at the infrastructure level as well as the personal hardware level. When we are truly free to move around and access data anywhere we want, there will be a huge explosion of uptake and new users, I think.
I'd better prepare the tin foil to head off my 60Ghz allergy.
Because wired has less problems. Wireless is nice in that, well, it doesn't require wires. So no cables to run. Less hassle in terms of physical effort, and you can move around while using it but that's where the advantages end. Wired has some big advantages:
1) Security. With wireless, there is always the issue of other people listening to your signal. Unless you live in a farady cage, you can't control where that signal goes. That means you have to deal with shit like encrypting the entire signal. That takes additional configuration to make work, and additional hardware to accomplish at high speeds. While AES isn't particularly intensive, try doing it at a gigabit. It'll hit a modern CPU hard and no way some cheap embedded device pulls it off without ASICs to help.
2) Contention. With a wireless system, you are all using the same bandwidth. This means it doesn't scale well with more connections. The more computers you have on it, the lower your total throughput. Not a problem with wired connections, each computer gets dedicated bandwidth to the switch. So I can transfer to you at full bandwidth while two other people also transfer at full bandwidth and there's no contention.
3) Range. Even under pretty good conditions, wireless doesn't match up to the distance you can get from a normal Cat-6 run (100 meters). Of course you also have wired technology for longer runs (like fiber), or you can simply have a switch repeat the signal.
4) Simplicity. While it is more work to lay the wires, once done you have less effort. A system just plugs in and all necessary information can be provided to it, no config necessary. With wireless, configuration must be done on the client machine, at least if any encryption is to be involved.
5) Reliability. Wireless just has problems. Be it interference from other devices on the same band, dead zones, weather, whatever, you can lose wireless signal because of too low a SNR. Not the case with a wired connection. They tend to always work, unless the cable breaks and that is quite rare.
6) Speed. Whatever you can do with wireless, you can probalby do better with wired. Just tends to be the case. This is particularly true if you include fiber in the wired category, but even if not. Right now N is as good as it gets wireless which gets maybe 100mbps of throughput max in terms of actual data (300mbps data rate, but there's tons of overhead). 1gpbs wired is common, 10gbps is available over regular twisted pair. Faster is being developed for normal twisted pair, and faster is already available for fiber or something like CX4.
Nothing wrong with wireless, but it is an addition to wired, not a replacement. I have a WAP so that I can use my laptop everywhere in my house. However my desktop, my Blu-ray, etc are all hard wired. I don't see that changing.
Am I correct in thinking that as the frequency of microwave radiation increases towards the infrared end (1THz), the radiation behaves more like infrared, i.e. impermeable through the thermal insulation of buildings? 60Ghz seems a big jump from the usual 2-5GHz for wifi.
"I bless every day that I continue to live, for every day is pure profit."
As I understand it, this is a replacement for running a fibre optic link between your house and your ISP. Instead, you mount an antenna on your roof, which engages in narrow beam, line of sight 60 GHz communication with your ISP. I think the benefits are that it is potentially cheaper than running a fibre optic cable to your house. The signal is attenuated by rain, and by atmospheric oxygen. I doubt the signal can travel very well through walls. And I don't think it is useful for mobile devices.
Doug Moen
I have written a truly remarkable program which this sig is too small to contain.
That the promise of 100Mb/s over WiFi really isn't realized i.e. 802.11n in every piece of equipment I've had my hands on delivers about 32Mb/s and in the same test with 100Mb Ethernet I get about 64Mb/s. It's kind of hard to take this seriously or that we simply have to take all pronouncements from the WiFi consortium at a severe "discount". Mind you if for some reason this difference was proportional to the other promises (and I can see no reason why it has to be). Even getting 1Gb/s over WiFi would be a drastic improvement. Somehow I doubt that this is the case though.
So the higher the frequency of your signal, the more bandwidth you can get. Easy to understand why. However there is a tradeoff, and that is distance/penetration. Low frequency signals can travel extremely long distances, and penetrate through material well. The ultimate example is the sub communication systems like Seafarer. That system, operating at a 76Hz carrier, could penetrate the entire Earth and send signals to submerged subs anywhere, at a rate of about 3 characters per minute.
So as you go up, the opposite is true. Go up to the 100s of GHz and you can carry astounding amounts of data if you like, but you find that the air itself will attenuate your signal a whole lot, and forget about a wall or the like.
This is why there's competition for various ranges of the spectrum, like 700MHz. One range is not as good as any other. Were that the case, we'd have no problem as there is plenty of space up in the high GHz range. However it's not. Low frequency spectrum can be very useful for things.
At 60GHz, you are going to need line of sight pretty much. It might penetrate a bit of stuff, but you can forget about having an access point 5 rooms over that goes through a few walls.
For a point-to-point outdoor link it'd work ok, though it would be the kind of thing that would suffer from reduced data rate or a completely dropped signal in the rain and rain plays hell on signals that high frequency.
So I can see it for special cases, but the next WiFi it will not be.
For most practical purposes, 60GHz signals don't penetrate anything. They just bounce around like light.
This stuff might be good for fixed point-to-point links, but that's about it.
I've worked a bit with existing 60GHz products, and while they're generally faster than greased shit, the alignment of them is typically very critical and, sometimes, even seasonal. This isn't the sort of product that would be useful for municipal wifi, except perhaps as a backhaul between 802.11 radios.
Of course, like any new product where there's money to be made, the marketers will claim that it slices, it dices, and it makes Julienne fries. Caveat emptor, etc. (But wait! There's more! If you act now, the sky will always be blue, you'll always be young, and you'll ejaculate rainbows.)
Meh.
Kid-proof tablet..
For my ansible
My internet connection ( DSL ) can't come anywhere near saturating my 802.11g router's 54Gbps. If my wireless connection is 10Mbps or 100Gbps, what does it matter? Unless I have a fiber optic line running to my home, how do I benefit from faster wireless? So at work I can open my TPS report off the local outlook server a fraction of a second more quickly?
I'm limited by the speed of my DSL, not the wireless connection speed.
"WiFi is simplex, meaning that it tx and rx half the time, so your bandwidth is necessarily less than half the marketing fluff figure."
-MiMo... heard of it?
That's a poorly chosen word, it makes me want to say wig-ig instead of wi-gig.
x86, oh yes, I'm pro.
...but you CAN see this.
WAAAOOOW!
Oh myyyyyy......
If the only way you can accept an assertion is by faith, then you are conceding that it can't be taken on its own merits
Consumer companies will jump on this shit like crazy just to maintain teh price point of wireless routers and APs. I always expected to get a 802.11g router for cheap once 802.11n came out. Instead, it's harder to find g routers.
To me, and most people I know, a new 802.11 standard won't mean a row of beans and yet they'll still have to shell out $50 to buy a new router when they spill their coffee on it.
52 52'23" W 47 32'07" N
I am surprised no one mentioned this in comments: the 60GHz band is unlicensed because it is impractical to use it. The Oxygen molecule has a resonance at 60GHz and highly attenuates most electromagnetic radiation around the 60GHz band. This means one wastes a lot of power while broadcasting around 60GHz, making the band pretty much useless. The English wikipedia article on 'Attenuation' has the image http://en.wikipedia.org/wiki/File:Micrwavattrp.png which shows a graph of attenuation in air vs. frequency. It is something like 0.008dB/km at 10GHz and something like 6dB/km at 60GHz.
I for one think this is not going to be practical.
We can't even get 10Gbps wired in the stores--not anywhere close to a reasonable price point--and that technology has been around for several years.
The only way this wireless technology gets into stores this year is if it functions essentially like the existing $700 paired HDTV transmitters. There's no multiple point-to-point, routing functionality and it's close range only.
Hatred of wires may not be possessed by Joe Average, but it is by Jane Average. In the living room, this counts. See also WAF.
maybe I'm being slow here, but none of my computers can cope with data coming in at 7gbps.
A couple of them have 1gbps NICs but the bottle neck for getting an actual sustained flow of data into the CPU is still the speed of the ISA bus (since none of the promised next generation of super fast buses ever actually appeared in any hardware)
what the hell protocol will you be using that has 1400% overheads? I think it's time to redesign the protocol in that case, rather than throwing more hardware at the problem :)
So the higher the frequency of your signal, the more bandwidth you can get.
I'm sorry, but this is just wrong. The higher the bandwidth of your signal, the more data (symbols viz. Nyquist) can be sent.
If one were to use Single Side Band transmission, for example, the carrier could be at 76Hz and, using simple Frequency Modulation this could be extended towards say 60GHz, giving, well a Freeking Large symbol rate.
Of course, doing this would interfere with every single electronic and probably every single electric item such as I dunno, a 60Hz light bulb. But it could be done
(You'll need to grep wikipedia for thing I should have linked to. Tired. Very tired)
--
Thanks fish, so long, or something.
So the higher the frequency of your signal, the more bandwidth you can get.
I'm sorry, but this is just wrong. The higher the bandwidth of your signal, the more data (symbols viz. Nyquist) can be sent.
If one were to use Single Side Band transmission, for example, the carrier could be at 76Hz and, using simple Amplitude Modulation this could be extended towards say 60GHz, giving, well a Freeking Large symbol rate.
Of course, doing this would interfere with every single electronic and probably every single electric item such as I dunno, a 60Hz light bulb. But it could be done, says Mr Tesla.
(You'll need to grep wikipedia for thing I should have linked to. Tired. Very tired. )
--
Brain the size of a peanut.
The above is corrected version courtesy of the cup of v. strong coffee on my desk, and is bought to you by the number 42 and the letter fark.
Yes and no.
There are limitations on how much can be transmitted at any given frequency, but a cellular type system of wi-fi could easily be used, probably at 60GHz. (I'm not sure how it is at penetrating walls though.)
Still, there are frequencies that would work. A low-powered cellular system where each repeater could cover, say, 10 blocks. Space them through a city with a 25% overlap, and a municipal network becomes quite plausible. It's no longer the nice, simple, thing that was originally evisioned, but it can work and can scale. Then you need wires to each of the repeaters, but that's a much smaller problem.
And the nice things about this "cellular" system is that the technologies needed have already been developed. But when you think of it as an internet system rather than a phone system, the constraints and options are different. With this system VOIP can really replace the phone...but you may still have delay problems. It's not designed as a real-time system. There's plenty of wavelength to have both systems, though. At worst you need to start placing your cell towers closer together.
OTOH!!! Phones tend to be designed to be usable in times of disaster. Internet and cell phone systems tend to go down at such times. These are design problems, but they are still real problems. It's cheaper to build them for only 97% up time, but ...
I think we've pushed this "anyone can grow up to be president" thing too far.
The problem is exactly that it can't possibly scale. There is only so much bandwidth physically available in each cell, and bandwidth demand isn't going to decrease any time soon. The only thing you can do is shrink the cells, and soon enough you're at the point where your cells are so small that you have to have wires everywhere anyway.
It's still useful, but it'll only be a slow backup for the real network that is using wires of some kind or other (copper, glass, whatever we come up with later).
I see two ways this could be a massive boon to the slashdotter community.
Fast external harddrives. 7gbps could make a wicked fast SAN in a shoebox. Since it would only need a power cable, cable routing would be a non issue.
This could also make a slick replacement for SATA if executed correctly. No more snaking SATA cables.
And the second, is an HDMI analog. No more stupid display wires? Awesome!
I'm not an electrical engineer, just some percolation from my mind.
So isn't 60GHz the same frequency that the new body scanner's (the one's that see your private area outlined through clothing) at Airports?
Is this the same technology that caused DNA to "unravel" on 7% of tests they ran?
Is this really a safe technology, or is this being rushed out the door for the sake of making a profit?
Unless you put APs everywhere so you have line of sight... but... meh.
I run: Windows, OS X, Linux, FreeBSD. Just because you have a hammer, doesn't mean everything is a nail.
It scales quite easily. If a cell gets too densly populated, you double the repeaters and drop the power.
Splitting the cell into four blocks is nicer than merely splitting it into two, but whether that's the reasonable approach depends on the population distribution.
Also if the city sprawls instead of packing, you build more repeaters further out.
I think we've pushed this "anyone can grow up to be president" thing too far.
But that is not scaling. It only works for a short while. Then your cells become too small to be useful, and scaling stops.
Why do you say it isn't scaling? It scales with the number of people.
Why do you say it stops working? It works clear down to each apartment having it's own repeater.
It only stops being useful if you insist that people subscribe to a particular cell. If it's being run by a municipality, then you can change from repeater to repeater without worrying about "can my system work on the new cell?" It's true that this is a bother if each cell requires a new contract of something, as in "You were in a GTE cell, but now you're in a Verizon cell", but there's no reason a municipal net should have that kind of problem.
FWIW, a municipality could put up cell towers for coverage much more cheaply than the telcos do, because there WOULDN'T be competition for coverage. There would a an attempt at total coverage. The telcos could rent "bandwidth" to run their service over the municipal net, rather like trucks run over the municipal streets. Just imagine what it would cost if each company had to lay it's own streets. (OK, it's not an accurate analogy, but despite ways in which it doesn't fit, it still points at inherent inefficiencies of the current system.)
I think we've pushed this "anyone can grow up to be president" thing too far.