10Gbps Wireless Transfers

Erasei writes "NTT Science and Core Technology Laboratory Group has developed a wireless communications that is capable of transmitting data at speeds of up to 10Gbps. In order to achieve such high data transmission speeds, the system uses the as-yet-unused 120GHz frequency band. The actual bandwidth the system uses is 17GHz, and the method of modulation employed is amplitude shift keying."

Comments

OK... keep us posted

[Gruneun]

In order to extend the range of the transmission, a 20cm-diameter lens was used to focus the beam. In the laboratory, a 10m range for the system has been confirmed, but NTT is expecting to try and extend the range to 100m

From a scientific standpoint, this is somewhat interesting. I don't see it being news, however, until they seriosuly decrease the size of the transmitters and/or increase the range. The transmitters look to be about the size of a shoebox, which is great for building-to-building, but let us know when they actually bump up the range to say... building-to-building distances.

It's Gbit/s, not Gbps -- And it's a big problem!

[SysKoll]

Journalists are fond of using dumbed-down abreviations such as Gbps or other acronyms. But I encourage technical writers to use the correct term, which is Gbit/s. Just as Mbit/s, Kbit/s.

There is something weird with computer science. People in this discipline badly need a common linguo because the field is evolving so fast. And yet, most CS practitionners couldn't be bothered to use the generally accepted vocabulary or abbreviations to describe their domain's problems. They invent their own, or incorrectly reuse existing jargon swiped from other disciplines.

Even worse, each branch of CS reuses general vague terms and overloads them with a different meaning. What's a "server"? What's a "page"? Depends who's talking.

As a result, CS is full of islands of disconnected knowledge and of specialists that cannot communicate with each other. Ever tried to have an OO programmer and a database admin talk to each other?

Mathematicians don't speak each other's linguo. But they carefully avoid using overlapping terms to define different things. That's what we should aim to do.

CS will keep being a cottage industry and a craftman discipline akin to voodoo, with cancelled project when wizardry fails, as long as all CS won't agree to speak a common language. Or at least a language where precision removes the overlapping meaning.

Granted, a precise vocabulary will not cure all the ills of CS. It not a sufficient condition for clear communication. But it's one of the necessary conditions.

So do your part. Write Gbit/s, not GBPS or other atrocities.

-- SysKoll

again, new tech but when will it be available?

[chamenos]

i've already mentioned this before and i don't wish to sound like a trolling wet-blanket, but when is this going to be widely available and adopted? slashdot is chockful of articles of new and emerging technologies that promise ever-increasing leaps and bounds in all these high-tech gadgets that all of us love (which is why we're here on slashdot!). most of these gadgets or other new developments in science and technology only see the light of exhibition or convetion halls and after that, either lousy manging, marketing or just a bad business model kills the product in its infant stages.

the article mentioned that this new wireless system uses a yet un-used 120GHz frequency. i personally feel that this can -both- be a disadvantage and advantage to its entry into the mainstream market. reason being, due to the 120GHz frequency being unused, widespread adoption might be made easier without any regulations but then again, that same lack of regulations would make it hard to regulate and control, and it might be open to abuses by users with malicious intent. worst still, all our beloved governments or telecommunication companies (for some countries, the above two are one and the same for all intents and purposes) will seize the opportunity to reap a handsome profit and end up killing the potential this product might have had in the market.

another factor, and perhaps the -most- important factor to consider is cost. the system may be cheap to built and maybe it didn't cost that much to develop, but we all know capitalism isn't about selling products at a price that indicates its real worth, capabalities, and cost of manufacture or R&D. rather, market demand or greedy corporate figures play an important part, along with sleazy marketing methods that target the ill-informed but rich people, who buy into meaningless numbers generated by the afore-mentioned marketing folks. the success or failure depends on how well the balance, between reaping the most profits and at the same time allowing growth, expansion and demand of the product to propagate is upheld by those corporate capitalist (and usually greedy) folks.

lastly, but not as importantly as the above factor, is whether the four times increase in speed is really needed, or is it on par with a 2.0GHz pentium 4, in that both only have impressive numbers to show off, which represent little of any real-world increase or -need- for performance for most mainstream users. but then again, i've just made a moot point haven't i? the sad reality is that the actual technology or the need for it isn't what sells. its how well its marketed (read: exaggerated and made pretty with big numbers) that really counts, ain't it?

if anyone feels like modding me down, go right ahead. i just had to get that off my mind. i just can't help being a little more and more pessimistic each time a promise of some new-fangled gadget makes headlines on slashdot, only to make headlines a year or two later when it flops over on its belly.

So why is IRDA so slow?

[BeBoxer]

That's what I want to know. IR is way up in the THz range. You outta be able to modulate a least a Gbps onto it easily. Sure, it wouldn't go very far but so what. Is it just the difficulty in making emitters and detecters which are fast enough? That can't really be that hard. A gige optical SX GBIC module can be had for under $100. Hm. That makes me want to tear a couple apart and see if I can get a link running across a few feet of free space.

Isn't Less More?

[mmol_6453]

Don't you generally want to send more data over less bandwidth?

Failing the "more data" point, I'd like to see a lot less bandwidth used, just to accomodate the artificial "crowding" of the spectrum. At this point, 10Gbps isn't very useful yet.

I mean, look, people! You're talking about 1.7 GHz/bit. To put that in comparison, think of extreme environments. If a deep-space probe developer only had a 17Hz signal, would you want to get only 1 bit per second, instead of the theoretically possible 8 bits per second? (IIRC, the theoretical limit for data transfer on a specific RF wavelength is freq/2. Any more, and you can't tell signal from noise without something as redundant as a UPS address stamp.)

I know this sounds like a "640k is all you'll need"-type-argument, but why not use a smaller bandwidth, and allow for multiple independant channels? If you need additional data transfer, you could use channel bonding.

This'll especially help wireless providers, who have to relay their signal from tower to tower without hitting the tower-after-the-next with a poorer signal. (Signals will travel beyond the horizon, but not very reliably.) Currently, they use polarization of their signal.