New wireless "WiFiber" tech faster than fiber?

Virginia startup GigaBeam set their sights high with their new wireless system: they're taking on fiber. They dub the tech "WiFiber," and are positioning the product as a replacement for fiber optic pipes in situations when they're too expensive, too harmful to the environment, or just too slow to deploy, and they're claiming one gigabit per second speeds to sweeten the deal. WiFiber operates in the 71-76, 81-86, and 92-95GHz frequencies to avoid interference and so that the signal is less impeded by light rain or fog than current high-speed wireless competitors. It also transmits in a tight beam to avoid overlap, but you still need a line of sight, so heavy rain can spoil your day. If conditions are right, though, GigaBeam's signal can make it for 10 miles, and is already shooting data from the Trump towers in NY and from a few other metro sites. Unfortunately, it's mainly for business links right now, since it's currently $30,000 for a set of radios (but that price should drop soon).

[Via Tech Review]

[CyBeR @ Feb 22nd 2006 2:33PM]

It does sound interesting (assuming the price drops quite a bit), but it is in no way faster than fiber. What they mean is faster to /deploy/ than fiber: you just set up the radios and you're done. With fiber (or any other kind of cable) it can take a while to get it laid down, not to mention the permits for digging, etc. Fiber optics now carry 40-gigabit streams on a single fiber pair.

[mike @ Feb 22nd 2006 2:51PM]

If line of site is required I'm not sure if it will ever take off in resendetial areas. But this whould work very well on skyscrapers as shown in the picture.

http://www.cranburypcrepair.com

[h2so4 @ Feb 22nd 2006 2:52PM]

I thought it was some kind of crazy monorail when I first saw the picture.

[Bancos Peru @ Feb 22nd 2006 3:03PM]

Looks amazing and it could be very useful to share bandwidth in a city:
1) WiFiber for the source and distribution points
2) Then normal WiFi for the final user

[Tracy L @ Feb 22nd 2006 3:03PM]

I thought skyscrapers were designed to "sway" a few feet from side to side, as opposed to being perfectly rigid. If true, would this interfere with reliability?

[furtim @ Feb 22nd 2006 3:06PM]

Oh, God. WiFiber? What a horrible, horrible name.

Cool idea, though! I remember reading about something very like this (in fact, probably the same thing) in an IEEE magazine last year some time. Except they didn't call it "WiFiber" then.

Good freaking God, is that a terrible name.

[DJ @ Feb 22nd 2006 3:14PM]

New dog, old tricks.

This has been tried before by several companies in the mid-90's, including Teligent, XO, WinStar and others. Identical concept and technology, just different piece of the spectrum. I worked for one of them for a while and while the concept was great, the business plans were fatally flawed. Spent money like drunken sailors! Maybe these guys can get it right, but I'm betting against it.

As far as residemtial, 3G will be easier and cheaper and here much quicker.

[Kaptain Korolev @ Feb 22nd 2006 3:22PM]

Is this new really new?

Yes the frequency of operation is increased but what this boils down to is essentially a microwave link. Look at your local cell phone tower in a built up area and you will most likely see what resembles a small drum attached to the antenna along with the cellular base station antennae. Microwave links are very very popular in telecomms for moving high volume raffic over line of sight links.

I can't see this being terribly reliable however. With a link using such a high frequency carrier the attentuation is going to be incredible if there is even the slightest hint of water vapour in the air.

[Bryan @ Feb 22nd 2006 3:39PM]

DJ, you're right. There was a company here in San Diego called AirFiber which ended up kaput, but I believe they made a few sales:

http://www.opticsreport.com/content/interview.php?interview_id=3016

[darkmoon @ Feb 22nd 2006 3:39PM]

Current fiber using DWDM is in terabit throughput.

[Lenny Markh @ Feb 22nd 2006 3:44PM]

Wouldn't this be best suited for a mesh network? Line of site would be easier to establish.

[Glynton @ Feb 22nd 2006 3:51PM]

Anybody else find the idea of corporate mischief involving a giant mirror hanging from a helicopter funny?

Glyn

[Topmounter @ Feb 22nd 2006 3:55PM]

You guys are right, this isn't "new tech", just "new branding" or "new packaging".



And that headline, "New wireless "WiFiber" tech faster than fiber?" is just plain misleading and has nothing to do with the body of text.

The body specifically points out that it is an application / environment specific alternative to traditional fiber optic cable.

Also, I don't see how you can compare this very specific technology to "Fiber"... "FIBER" IS JUST A WIRE, that is like saying "OMG CAT-5 IS SO FAST", what determines its capability is how you deliver data over that wire.

[Jason @ Feb 22nd 2006 4:31PM]

A couple of years ago I read that Lucent got a DWDM system up to 72 TB/s over 3500km (unrepeatered).

[sines~ @ Feb 22nd 2006 6:11PM]

how much would $30,000 worth of fiber buy you? Would I be correct in assuming much of the cost goes into the deployment and laying of the cable rather than the fiber itself?

[kelly @ Feb 22nd 2006 10:06PM]

the cost goes into lighting the fiber up, not the deployment. jason, price out that 72TB interface. good luck, it doesn't exist. OC-48 cards alone in a cisco environment are $50k plus on one side (rough street price unless you're a huge customer), not counting the chassis/switching fabric costs. gig-e ports are relatively cheap. this company is basically saying they can save you the capex of lighting the fiber in a traditional manner, hand off a gig-e interface on the backside in a metro environment. $30k for the pair of radios to do it is way below cost of traditional WAN methods right now doing it yourself. someone that has a dwdm system and embedded capex may beat it, but lighting fiber has huge costs. this is why there's a ton of unlighted strands out there... all the telcos laid fiber during the boom, they lighted a fraction of them because of the cost.

this is of course assuming they can't do native gig-e across private fiber. cheap way to do it is with SFP gig-e ports with less expensive switches. cheap, easy, but you need dedicated fiber pt to pt.

but I may be wrong :) been awhile since I was on the techie side of things.

[Jo @ Feb 23rd 2006 5:45AM]

GB fiber speeds should be cheap in densely populated cities due to the economics of scale (ISP permitting). That�s why cities such as Hong Kong have cheap, fast broadband. This would be news if it would cover rural or suburban areas with GB speeds.

[Ricardo Taveira @ Feb 23rd 2006 11:54AM]

Wow, this is great, one of my favorite websites covering my company.

Here are a few issues brought up in the comments:

1)Faster than fiber is misleading. That's not what we're about. Nobody's dreaming of having a 72TB link yet. We have over 40 1.25GBps links deployed, including an 8-mile link. We'll have a 2.7GBps by Q2 and we'll get OC-192 and 10Gigabit Ethernet by the end of the year.

2)This is new. New spectrum, new capabilities, new company. Teligent and Winstar tried to create networks of their own. We sell the links, not the network.

2)Nope, it's reliable all right. We guarantee 99.999% reliability (that's a downtime of only 5 minutes a year!) over a mile for 80% of the US. We don't get interference from snow, dust, fog, etc. We only get intereference in TORRENTIAL rain--more than 3 inches AN HOUR. The other chunk of spectrum which tries something similar is 60GHz, and yes, they have tons of oxygen absorption problems (or benefits, if you dig the spin).

[Jason @ Feb 25th 2006 7:47PM]

Just as a side note here, the article says that "WiFiber operates in the 71-76, 81-86, and 92-95GHz frequencies to avoid interference and so that the signal is less impeded by light rain or fog than current high-speed wireless competitors." Yes, the higher frequencies will help avoid interference, and NO, they WILL NOT be less impeded by light rain or fog than current high-speed competitors - the higher the frequency, the more attenuation obstacles (rain, snow, fog, trees, buildings, etc.) will have upon the signal.

[Todd Laff @ Feb 28th 2006 9:47PM]

Simply Point to point wireless supports, or backhaul's to a wireless mesh network. These mesh networks are prone to interference and not able to produce 1GB or in many cases 100MB (it gets split each way). The advantage of 60GHz is no interference and able to deliver a true gigebit each way (but only up to 1.5 miles). 70/80GHz, like Bridgewave or Gigabeam can deliver a Gigebit over 3 miles.

This is good campus, building to building and backhaul to a wireless WAN.

[sunil @ Mar 14th 2006 4:23AM]

sir,
please give much information about " wifiber".

[unmesh @ Sep 18th 2008 6:40AM]

that itz ,it is nice thing .
please send me much innformation on my email

[rachanahrp @ Mar 11th 2009 3:36AM]

Reading the above comments, i feel that wifiber is again a simple wireless tech same as wifi or wimax.... plz may knw that wat i think is true or false??
or if wifiber is really better than any other wireless communication, than may i knw hw??

[macintoshi @ Apr 29th 2009 7:06AM]

Please compare Wi-Fiber to Wi-Max in ter of speed and area wide range....