Ubiquitous Multi-Gigabit Wireless Within Three Years

Anonymous Howard passed us a link to the Press Escape blog, and a post about the future of ultra-fast wireless connectivity. Georgia Tech researchers unveiled plans to use ultra-high frequency radio transmissions to achieve very high data transmission rates over short distances. In a few years, the article says, we'll have ubiquitous multi-gigabit wireless connectivity, with some significant advances already under their belts. "GEDC team have already achieved wireless data-transfer rates of 15 gigabits per second (Gbps) at a distance of 1 meter, 10 Gbps at 2 meters and 5 Gbps at 5 meters. 'The goal here is to maximize data throughput to make possible a host of new wireless applications for home and office connectivity,' said Prof. Joy Laskar, GEDC director and lead researcher on the project along with Stephane Pinel. Pinel is confident that Very high speed, p2p data connections could be available potentially in less than two years. The research could lead to devices such as external hard drives, laptop computers, MP-3 players, cell phones, commercial kiosks and others could transfer huge amounts of data in seconds while data centers could install racks of servers without the customary jumble of wires."

Microwaving your privates?

[Aqua_boy17 (962670)]

we'll have ubiquitous multi-gigabit wireless connectivity, with some significant advances already under their belts.

If they're running this from laptops for extended periods, that may be the only thing remaining under their belts.

Re:Microwaving your privates?

[Moby Cock (771358)]

Technically speaking, at 60GHz, you'd be millimetrewaving your privates.

...for that matter...

[drakaan (688386)]

while data centers could install racks of servers without the customary jumble of wires

Somehow I don't see "whole data centers" using a data transmission method where any device can potentially intercept the data going to and coming from any other device. Might make your hosting clients a bit nervous.

Re:Microwaving your privates?

[BlueParrot (965239)]

It doesn't make any sense to make a network card emit microwaves at intensities similar to microwaves because not only would you get a huge power consumption, it is also massive overkill unless you plan to search the sky for stelth bombers. The FCC ( or local equivalent ) would probably have a few things to say about it as well. The scaremongering about radiation from comunications equipment is simple unbeleivable. You are more likely to get hurt from tripping in a cat5e cable.

I am a data center manager

[Travoltus (110240)]

Maybe some lower security data centers might enable wireless, but I doubt it. Being that we're a financial institution (a small one, mind you), there's no way in the h to the e to the double hockey sticks that I'd ever enable any kind of wireless anything in our data center.

I'd rather deal with a network cable gone sentient and whipping around like a snake and attacking people, than go wireless at the data center.

Only an idiot thinks there's a wireless transmission that's invulnerable to being intercepted. Heck, wired communications aren't 100% secure, either, but my boss's business is about minimizing risk, and wireless networks even inside a data center is not minimizing risk.

Re:

[tedrlord (95173)]

I figure you could put it in a faraday cage of some sort. Still, I'd prefer a little planning and cable management to several hundred machines and peripherals transmitting wirelessly anyday. Especially since I have to spend days on end in there every so often.

Re:

[markov_chain (202465)]

Faraday cage wouldn't work. Any openings need to be smaller than the depth of the skin current, or the signal induced on the inside surface will just flow out through the cracks and re-radiate.

It would suck anyway

[Wesley Felter (138342)]

Sure you can get 15Gbps, but if you start sharing that bandwidth among dozens of servers it wouldn't be all that fast anyway.

Re:I am a data center manager

[walt-sjc (145127)]

My little cage at the colo doesn't have 5 servers. It has hundreds. I'm also sharing that datacenter with many many other companies that have cages with hundreds of servers. We deal with SAN / iSCSI, NAS, backups over networks, etc. With the noise and limited bandwidth available in a shared frequency space, I seriously doubt any type of wireless will be very useful in a datacenter - especially since everything is already connected via hard-wired connections.

It also won't be very useful in my home, where wires are already easy to run for the short-distance devices, and noise / distance prohibits the use in cases where I could really use and WANT high-speed wireless.

So it does sound like a neat trick, but what is a valid, viable use case for it?

I could REALLY use something much different. I want to get rid of the 20 or so wall-wart power supplies under my desk. I want one larger power supply that I can run small cables to all the devices. Why can't devices negotiate for how much voltage / current they need?

Re:

[The Cisco Kid (31490)]

I want something different too. I dont want higher speed, I want more range. I want one or two megabits at 30 miles NLOS. Either simple point-to-point, with many different 'channels' for seperation, or point-to-multipoint. Of course, the question of wether such a thing is technically possible is irrelevant, because the telco's would kill it in its crib anyway.

Re:

[phallstrom (69697)]

> So it does sound like a neat trick, but what is a valid, viable use case for it?

Maybe for your AV stuff? No wires b/n your dvd, receiver, and tv would be nice.

Re:

[drinkypoo (153816)]

My little cage at the colo doesn't have 5 servers. It has hundreds. I'm also sharing that datacenter with many many other companies that have cages with hundreds of servers. We deal with SAN / iSCSI, NAS, backups over networks, etc. With the noise and limited bandwidth available in a shared frequency space, I seriously doubt any type of wireless will be very useful in a datacenter - especially since everything is already connected via hard-wired connections.

I've seen security rooms inside datacenters that had copper cloth over the windows, etc etc. What if every cage in the colo were a faraday cage? In theory, wouldn't that permit this? Or, how about UWB? Isn't UWB supposed to allow an effectively infinite number of transmitter/receiver pairs to operate together? If the whole building were shielded so that it wouldn't penetrate, it would eliminate interference issues.

I still think that fiber is more desirable. I wish it were cheaper (although it's getting c

Re:

[interstellar_donkey (200782)]

That's exactly why, when I first read about this, that I thought that the appeal of high speed wireless would mostly be on the consumer end. Most businesses are bound to see the potential security risk of wireless and stay away from it, regardless of how fast it is.

As I don't manage any data centers, I'd love it. Mostly because the wife has forbade me from running CAT5 through the house and I'm stuck with 802.11g connections. It's annoying to try to transfer a large file from the office upstairs to, say,

Silly Fears.

[twitter (104583)]

there's no way in the h to the e to the double hockey sticks that I'd ever enable any kind of wireless anything in our data center. ... my boss's business is about minimizing risk, and wireless networks even inside a data center is not minimizing risk.

Your network is on the internet. That and any non free software you have are bigger threats than sftp over wireless.

Re:

[Macthorpe (960048)]

That and any non free software you have are bigger threats than sftp over wireless.

Going to cite anything for this one? Or am I going to be left waiting like I have with this gem? [slashdot.org]

Quit now

[Travoltus (110240)]

before you get caught up in a security breach scandal and the orders they gave you to implement wireless networking get sealed up in one of Dubya's supersecret war on terror files.

Polish up your resume and quit now.

Really, I'm not kidding.

FTFA

[SighKoPath (956085)]

Pinel is quick to point out that a multi-gigabit wireless system would present no health concerns as the transmitted power is extremely low, in the vicinity of 10 milliwatts or less and the 60 GHz frequency is stopped by human skin and cannot penetrate the body. The team admits that the fact that multi-gigabit transmission is easily stopped means that line-of-sight is essential, and this could be a stumbling block in practical settings.

Doesn't this make it being wireless kinda pointless? It's like a wired connection where you can't step over the cable or drill a hole through the wall!

Re:

[MontyApollo (849862)]

You could wire a transmitter/access point into everyroom near the lights. You would still have to wire indoors, but you would have untethered movement.

Re:

[MontyApollo (849862)]

Or you could just put relays in line of sight of one another. You wouldn't need too many if each device also could relay.

Re:FTFA

[Moby Cock (771358)]

Useless? No. But very application specific. However, there is a great appeal in making Personal Area Networks.

That and being able to connect a DVD player to a TV without a cable would be, in a purely geek way, quite elegant.

Re:

[gatzke (2977)]

Until someone turns on a microwave.

Or you live in an apartment and your n nearest neighbors compete for bandwidth.

Or somebody nukes us and the EMP keeps you from watching American Idol.

Remote display and input

[MontyApollo (849862)]

Could this kind of bandwidth run a remote display?

I always thought it would be cool to have a pad that was nothing more than a screen and input device that you could carry around the home instead of a full-fledged laptop. You would be actually "running" your powerful desktop off basically a second screen that you could carry around with you in the house.

Re:

[Moby Cock (771358)]

Didn't Capt Picard have one of those?

Re:

[Yosho (135835)]

Well, let's do some math. Let's say we've got a 1680x1050 display at 24 bpp and an update rate of 60 Hz. That's 1680*1050*24*60 bits per second -- in other words, 2.37 Gbps. So, yes, a connection like this could conceivably run a remote display.

Great!

[MightyMartian (840721)]

You'll be able to watch pr0n through your neighbors open wireless network *and* fry up a steak by positioning the frying pan between the access point and your notebook. Don't worry, the sunburn should fade in a few weeks.

Not for the data center

[nincehelser (935936)]

I can't see any real application for this in a data center. They'll always use wires, switches, and routers. One simple reason is that one bad wireless transmitter could jam a whole bunch of nearby servers, which probably wouldn't be good. Wires have their uses. Sometimes it's good to keep your data flow contained and controlled.

ubiqutous, multi gigabit pornography

[by Anonymous Coward]

great. now ill never have a reason to meet girls

2 ways to increase thruput

[ookabooka (731013)]

There are 2 ways to increase the amount of data that can be sent. Increase the carrier frequency or increase the bandwidth. What these people have done is increase the carrier frequency. Wireless today runs on 2.4ghz, these devices run up to 60ghz. What does that mean? Well it'll take more energy, higher frequency means higher energy, also it attenuates more, meaning shorter range. Not only that, but it can will be more readily absorbed by things like bricks, desks, your foot, etc.

The alternative to this is to increase bandwidth, say use 2.1ghz through 2.6ghz for 1 signal. The obviously downsides to this are you can't run many concurrent streams.
All in all wireless data transfer has a very real ceiling on the amount of data that can be transferred, lower frequency means longer range and ability to go through obstacles, at the cost of reduced data-carrying capacity. I guess the point of this post is to point out that there is only so far we can go with wireless data transfer. I don't think it will be able to keep up (over the long run) with the increasing size of traffic to be a viable alternative to cables when it comes to things like comptuer networking. Anyone have any thoughts on this?

Re:

[Moby Cock (771358)]

The use of multiplexing codes has not been fully exploited, yet. MIMO and others are used extensively in cellular networks (which are, let's face it, wireless networks too) but are less common in 802.11 and similar networks.

Perhaps the next generation of wireless will include UWB/CDMA based transmission.

Re:

[ookabooka (731013)]

Even with multiplexing there is still a very real limit to the throughput of a certain frequency. I suppose my point is that there are clever ways to allocate bandwidth to users depending on how much they need, or to combine a bunch of frequencies to get the throughput you need, but it just isn't realistic to think that one day everything can be wireless and sending movies to and from each other no problem. Basically with wires you can do intelligent switching, but wireless requires you to broadcast and tak

Re:

[Detritus (11846)]

The relevant parameters are bandwidth and signal-to-noise ratio, not the carrier frequency. See the Shannon-Hartley theorem [wikipedia.org] for details.

Re:2 ways to increase thruput

[rcw-work (30090)]

There are 2 ways to increase the amount of data that can be sent.

There are actually four:

• Increase the signal strength (using a directional antenna or amplifier)
• Decrease noise (use higher-quality components, shut off interfering transmitters, use directional antennas)
• Increase the signal bandwidth
• Increase signal spectral efficiency (for example use OFDM instead of FSK)

Changing the carrier frequency has no effect, except that there's more room for higher-bandwidth signals at higher frequencies. 2.400-2.422GHz seems like a smaller chunk than 400-422MHz, but it can carry the same data.

The formula for how many bits you can send and receive error-free is the Shannon-Hartley theorem [wikipedia.org], and spectral efficiency is typically stated as a percentage of the theoretical.

Interesting technology

[retro128 (318602)]

This technology could be used in applications besides just strict data transfer. 15Gbs should be fast enough to drive a display, as well. The proverbial rats' nest behind your computer could completely disappear with this technology. Keyboards, mice, displays, network - Just about cable plugged into the back of your computer could be replaced with wireless this fast.

But if only it were so simple. Of course now the problem we have is with security. Never mind TEMPEST [wikipedia.org]. If you had a big enough antenna and you could decrypt (it IS encrypted...heavily...right?) the datastream emanating from this technology from a distance - you could see the display, keystrokes, data transfers, everything. Obviously, strong encryption is very important - But the overhead from strong encryption will reduce the theoretical bandwidth because of the extra baggage on the packets, and increase costs significantly because of the very specialized ASICs that will likely be required to encrypt a stream at that speed. And they'd have to be standard across all devices. AND an exploit had better not be discovered in the algorithm. Then there's the issue of the 60GHz band. A frequency that high is very unforgiving of obstructions, even at the short ranges we're talking about. If you have a metal desk, forget it. And what about jamming from computers in close proximity? What about from a "l33t hax0r" with some time on his hands and an inclination to make trouble?

Re:

[drinkypoo (153816)]

If you had a big enough antenna and you could decrypt (it IS encrypted...heavily...right?) the datastream emanating from this technology from a distance - you could see the display, keystrokes, data transfers, everything. Obviously, strong encryption is very important - But the overhead from strong encryption will reduce the theoretical bandwidth because of the extra baggage on the packets, and increase costs significantly because of the very specialized ASICs that will likely be required to encrypt a stream at that speed.

The real problem here therefore is one of cost. You can have as much bandwidth as you can pay for (because this is the kind of problem that responds well to parallelism. The penalty for that parallelism need not be all that significant. You can have no encryption cheaply, but uh, yeah. Next.

I don't suppose anyone out there knows of any properties of physics that would allow for linked "random" number generating systems that were consistent? :)

Re:

[RajivSLK (398494)]

The proverbial rats' nest behind your computer could completely disappear with this technology.

No, the problem you will then have is power. Everything still needs power. Keyboard, monitor, mouse etc.

Bull

[Conspiracy_Of_Doves (236787)]

It might be commercially 'possible' in a few years, and I'm sure that countries other than the US will even have it, but the US ISP monopolies will never make it available.

I think the summary went off the deep end..

[Kjella (173770)]

...when it said wireless in the data center. Yes, I've heard the theoretical figures for wi-fi. Try dropping a bunch of access points and various clients in tight proximity and see what it's really like. In a datacenter you can run 10x 10Gbps wires right next to eachother without problems. Can you do that with wireless? Hell no. I imagine the speeds quoted are ideal with free line-of-sight and no interference, good luck trying to achieve that in that bunch of wires. Personally I was fed up with wireless when I realized one AP couldn't even cover the ground floor of my parent's house. It'd take probably three to cover the whole house. Great... not.

Why this is partially a crock

[postbigbang (761081)]

The i-squared-r law at 60ghz means that even if the spectra was available (it's not) then you'll need both line of site (reflections won't help and will slow the data rate considerably) and you'll need the will to gulp content that fast. Of course, a shared fixture like an access point in WiFi suffers from duty-cycle problems and raw bandwidth will help. But we could also use spread-spectrum and/or advanced coding techniques like n-Pole modulation to accomplish the same thing.

Therefore, with all due respect

How about using the wired bandwidth first?

[Harald Paulsen (621759)]

The research could lead to devices such as external hard drives, laptop computers, MP-3 players, cell phones, commercial kiosks and others could transfer huge amounts of data in seconds

How about enabling my external USB-drive to use the 480Mbit available first? Or what about a NAS that can fill up a 1GBps ethernet? Wired isn't slow, it's just not used right.

Marketing lies as usual

[gweihir (88907)]

First, successful lab demonstration of multi-gigabyte speeds with mass-market capable technology is still missing. Call that at least 5 years to a real product. Then deployment. Who needs this stuff enough to deploy it immediately? Right, allmost nobody. Also the first product generation will not really be usable. Call it another 5 years to wide-scale deployment. That gives me an estimate of at least 10 years, but more likely 20 years. The 3 years are a direct lie, plain and simple.

I hope these ethically ch

Useless

[hcdejong (561314)]

I expect the line-of-sight requirement is a dealbreaker for 'personal area network' type situations. I've got my computer underneath my desk, and all gadgets that could possibly benefit from high-speed wireless links are above the desk. Reconfiguring my desk to provide LOS for everything (including keeping the desk clean, no stacks of paper between the computer and the gadgets) would be a major PITA. I'll stick with wired connections, thank you.

High-speed wireless could be useful for 'last mile' connections

Why not get a Government grant to research it?

[jd (1658)]

The Government spends money on everything else, why not spend it on something useful?

Re:Call me a luddite, but...

[by Anonymous Coward]

Yeah! Cause the biosphere wasn't already inundated with electromagnetic radiation. Its a good thing the rest of the universe doesn't spew loads of it towards the Earth. Oh wait...

Yes, and 99% of all CO2 on the earth is natural

[tjstork (137384)]

99% of all the CO2 in the atmosphere is natural, and we chalk up a change in climate to our 1% fluctuation, as if, that vast lion of 99% doesn't fluctuate on its own. So, why not worry about radiowaves in a radioactive universe.

Re:Call me a luddite, but...

[orclevegam (940336)]

but that's not really the same as saying that we will now saturate the biosphere with radiation of our own making.

As opposed to all that radiation saturating the biosphere not of our own making? You do realise that light is radiation right? Also, in case you're worried about all the terrible WiFi access points, your average 60 watt bulb puts off far more energy (radiation) than any WiFi AP in use. Now, admittedly, not all radiation has the same effect on everything (such as UV), but the key thing with EM radiation like light and radio waves is the total power and the distance from the source. Remember, power dissipates with the square of the distance, so if you're anything but sitting on top of the transmitter, and even then if it's relatively low power, you've got more to worry about standing outside on a sunny day. The fact that they're talking about such short distances with this tech leads me to believe this will probably be a very low power device, much the same as bluetooth and RFID are.

Re:

[by Anonymous Coward]

My concern is that we lack the science to even understand the implications of all of this radiation we're creating upon our environment. Sure, you can put a frog in a box next to a wireless system and say, "oh, the frog lived", or jack up the energy by 100 times as some sort of a proxy for exposure over time, and say "the frog did not get cancer", but that's not really the same as saying that we will now saturate the biosphere with radiation of our own making.

we understand electromagnetic radiation in grea

Re:Call me a luddite, but...

[physicsnick (1031656)]

UHF frequencies (millimeter waves and microwaves) cannot cause cancer. The photon energy is not high enough to break chemical bonds in biological tissue.

When a chemical bond is formed (say, in DNA), a certain amount of energy is released. To break that bond (and cause cancer), you need to put that energy back. The catch is, because of quantum mechanics, the energy can't be accumulated. You can't pile in more and more photons until it finally snaps; you have to get one big photon to come in and snap it. When you state the frequency of a photon source (e.g. 60 GHz), that indicates the energy of each individual photon (0.00024 eV). Typical bonds in DNA are on the order of hundreds of eV. It's physically impossible for this to cause cancer.

Even if you put your cat in a microwave oven, it won't get cancer (though it will die a pretty horrible death).

The danger with electromagnetic waves is heat and depth. UHF electromagnetic waves have far less energy per photon than visible light (~2.5 eV), but they have much greater depth penetration. They go deeper before they collide with your molecules, so they deposit heat deeper into your flesh than visible light or UV radiation. This is why putting your cat in a microwave is very bad; it essentially gets "cooked from the inside out". But the energy outputted by wireless devices is barely enough to cause even measurable changes in the temperature of human flesh. How much heat can you apply to a glass of water with a 1.5 V AA battery? Not much. Now spread that out spherically in a 100 meter radius. Almost zero.

Even then, biological organisms are very good at regulating their temperature; humans live across a wide variety of climates all across Earth, and yet still manage to balance their internal temperature.

Hence, UHF communications are not dangerous.

Re:

[drinkypoo (153816)]

I have this weird feeling that pervasive, high frequency radio needed to make wireless work is going to wind up with some unforseen bad side effect, the same way every other technology that we used too much had.

Even if for no reason other than having security of communications, it would be preferable if data were communicated via fiberoptic cable. Bonus points for creating optical transceivers that don't broadcast their signals all over the RF spectrum as a side-effect of operation.

Re:metric?

[by Anonymous Coward]

Says some who's obviously not old enough to drive a car and see "mph" or "kph" on the dash.

Posted anon cuz you ain't worth the karma. ;-)

Re:do not underestimate...

[Doctor Memory (6336)]

Maybe your friend's in the back yard, and you're on the second floor. Oh, and the RIAA's at the front door...