Free Nationwide Wireless Internet Access? 350
LiquidEdge writes "ISP-Planet is reporting that startup M2Z wants to offer 95% of America free wireless Internet access using the 20Mhz frequency allocation. They're backed by Kleiner Perkins, one of the most successful VC firms in history, and being started by the guy who built the @Home network and a former FCC Wireless Bureau Chief. 384/128 speeds will be free and they'll sell the higher speeds and the government will get a kickback of the revenue."
How do we make money? (Score:5, Interesting)
Ya, I know, just the low speed is free. But still, doesn't sound like a solid business plan. From what I understand, what people like most about broadband is the "always on" aspect.. not so much the bandwidth. I wonder if 384/128 is low enough to encourage people to pay for the faster service.
-matthew
Re:How do we make money? (Score:3, Interesting)
Generally I would tend to agree with you. IMHO, 128/128 would be a better bandwidth point. However, with the rise of Hi-Def media and multimedia over the Internet, it's only a matter of time before Hi-Def over the Internet becomes the standard. When that happens, I imagine that you'll see a lot of users looking for more bandwidth to power their in-SUV televisions/radios, their video-conferencing cellphones, and their on-the-spot videoblog reporting. All these concepts need is a network.
Err, on second thought, that's a rather scary thought.
Re:How do we make money? (Score:5, Interesting)
What stops me from getting 20 free wireless hookups and running a shotgun program to effectively combine the bandwidth? Other than some sort of account creation requirements (one connection per address? or per Credit card?) I don't see how they could really prevent this.
Sounds like a future OSS project if this project goes through
Re:A kickback on "free"? (Score:1, Interesting)
1) Gain monopoly control of a bandwidth portion, by promising to everyone that part of it will be "free."
2) As, inevitably, 384kbps speeds become entirely obsolete, the free portion of the service will no longer be worth anything to anybody.
3) Continue doing business, now effectively owning part of the spectrum and paying practically nothing in return.
(I can't read the article itself because it's Slashdotted; if their plan includes automatically scaling the free bandwidth as some percentage of their commercial bandwidth, then I retract my cynicism.)
More LEO and other trash; slower than EVDO, too (Score:3, Interesting)
Still another plan that will fail out of the starting gate. How about blimps, covering the horizon? 384k is barely usable. If you want it today, get an EV-DO card from Verizon or Sprint... or maybe an Edge card from Cingular/T-Mobile downstream-- once they can cover more than a few sq mi at a time.
This is not only money down a rat hole, but the announcement is also designed to queer all of the WiFi providers trying to build business cases across the country.
Not going to happen. Worse, it's obfuscation at its pinnacle.
Re:Not enough bandwidth (Score:1, Interesting)
Carrier frequency and bandiwith are two different things. Most of the time the first is much higher than the latter.
Not going to work the way you think it will.. (Score:3, Interesting)
Well, that would sort-of work. You might be able to get multiple transmitters to send on each one of the frequency slots on a given channel set. The largest problem to that really working is that you're going to have a hell of a time getting the transponders to sync up nicely and not collide and interfere with each other. The second problem is that "shotgunning" worked mostly because you were using multiple independent channels (Seperate phone lines...) and hooked in at the lowest device driver levels and aggregated the total bandwidth in a just so way so that parts of packets could be sent down one wire and other parts down another. You have only ONE channel and you're not going to very likely get the level of device access that you had with the dialup modems. You could get BGP to probably handle all of that if it's exposed only as an network type connection to the user, but since you're a freebie account, you're probably not going to get BGP to readily work because their routers won't acknowlege your router on the recieving end of the multiple recievers.
Well, the above items cause their own set of problems- but there's one more. If you succeed in doing this, you're guilty of theft of service- which is a felony offense in pretty much all of the US as a whole. If you encourage this practice, you're inciting to commit a crime- also a criminal offense in most states. Sure, you're not as likely to get caught doing it as with some other things, but I'm not one for commiting felonies just to get bandwith. Perhaps you don't have those moral qualms?
A common conversation (Score:4, Interesting)
customer: I just bought a laptop and the wireless internet stuff only works in my apartment.
me: Do you have an account with ?
customer: I don't need one. It's free here in .
me: Sorry, you're not an customer. There is no wireless internet available where you are.
customer: Yes there is! Flip over the other card and read that. *duh*
me:
customer: All new laptops come with free internet.
me: Great, but you still need to contact the ISP that your laptop is partnered with and sign up.
customer: You must be new, or something. You obviously don't get it. I just start up my laptop, and it says "Successfully connected to the Linksie System thingy" and off i go!
Re:95 percent? (Score:4, Interesting)
20MHz is allocated as Government/Non-Government Shared Fixed (Primary), Government Mobile (Secondary). Right next to that at 21MHz is 450kHz of Amateur Radio allocation (the 15-meter band). You can do long-distance (DX) communications on 15-meter, including around the world, if conditions are right.
In other words, with an allocation in the 20MHz range, a user is theoretically capable of covering an entire continent with just a single tower. Even if they use a relatively small number of towers (which would be realistically required, if not theoretically), all of the towers will most likely be able to at least interfere with each other. So they'll all be using the same frequencies, and therefore all sharing the same bandwidth, even if there are multiple towers. This is unlike cell towers. Two neighboring cell towers may overlap, but a cell tower 50 miles away can use the same frequency (and therefore bandwidth) without a problem.
So, how much bandwidth will they be able to provide? Let's assume a fairly high-tech encoding: 64-QAM or OFDM. Nyquist tell us that bandwidth = 2 * bandwidth * Log2 (states/signal) bits/second, or 2 * 1MHz * Log2(6), which is 3.6Mbit. For the entire area served by each tower.
But remember that these towers will cover a huge area. States, easily, and the entire continent regularly. I live in the Detroit, MI area. That's 6 Million people. That's about *half* a bit per second per person... And that's with a high-tech encoding like 64QAM.
Now I am not an EE, so please check my math. And I haven't read the article (only the summary), so if it's a 20MHz allocation in some other region of the spectrum (instead of a 1MHz allocation at 20MHz), then the story changes. However, even then, it's not great. You're most likely going to be limited to line-of-sight frequencies (the DX frequencies are already taken).
So, if it's an allocation of DX-capable frequencies at 20MHz, you can get away with a few towers, but you won't have enough bandwidth. And even if it's a dedicated 20MHz allocation somewhere else, you're going to need a bunch of towers.
What is the advantage of this over something like 802.11? I just don't know. No matter what, it seems like you'll need a number of towers comparable to cell phones today, even with a dedicated 20MHz of frequency. 802.11g uses 20MHz channels to provide 54Mbit of bandwidth using OFDM. So even assuming that the entire 20MHz is allocated exclusively to them (so it's cleaner than the ISM bands 802.11 works in), you're still only going to have 54Mbit of bandwidth (and likely only half that usable bandwidth) for your users. At 384kbps/user, you're looking at a theoretical maximum of 140 users per sector per tower, and a likely limit of 70. That's comparable to cell phone towers (roughly 100 users per sector).
In any case, this does not seem like a brilliant flash of inspiration in bringing broadband to the masses. It sounds like an attempt to create a government-backed monopoly on wireless communication. At least the cell companies had to buy their frequencies. In the end, I can't see the difference between this and digital cell service...
Getting started on your own large-scale network (Score:3, Interesting)
If you wanted to use a familiar brand, Cisco's Aironet 1300, http://www.cisco.com/en/US/products/ps5861/product s_configuration_guide_chapter09186a008021e5ca.html [cisco.com],
looks like another option except it costs ten times as much and I'm not
sure what advantages if any it has over the aforementioned device other
than perhaps the support you'd get from a larger company like Cisco. When
you deploy a network on such a scale, you're going to get people who use
it to download movie after movie, so advanced bandwidth throttling
(prioritizing certain types of traffic over others) would be key, and you
might have to pay up for something like this Cisco device for the traffic
shaping. Not sure about that...
For mega long range antennae to scatter around the neighborhood, as with the city of Cleveland which went wireless, have a look at this to learn more about the WISP (wireless internet service provider) deployment and equipment you'd need: http://www.trangobroadband.com/products/atlas_ptp. shtml [trangobroadband.com].
That company sells products that can beam twenty miles (line of sight, of course).
Spectrum Cash (Score:3, Interesting)
One of the secrets to the Clinton Administration's projected budget surplus(es) was they expected a lot of cash to come in from future spectrum auctions.
The specific auction (I think) that you're talking about was for the 2500 MHz to 2690 MHz band & they were planning to auction it off "no later" than 9/30/2002. There was a lot of problems with the plan, partially because the military uses a lot of those frequencies.
Anyways, Clinton was expecting that there would be big bucks made when the FCC auctioned off the TV spectrum after they switched over from analog to digital broadcast.
If you haven't noticed, that switch never happened, the FCC never got to auction off those frequencies, and the next President didn't have all that extra cash to play with.
I'm not blaming/defending either President, just pointing out that the future Clinton surpluses were heavily dependant on FCC spectrum auctions.
Re:Not enough bandwidth (Score:3, Interesting)
It's an interesting band as well. I'd love to see how they deal with it during more active cycles. Your 1 MHz slice might suddenly propegate for a few hundred miles - not exactly the kind of frequency you want for cell-based coverage (unless that is their plan - to only use a couple of nodes per state for "384" divided by tens of thousands of customers. I've been working 17 MHz an increasing amount as the solar cycle begins to wake up a little bit - from the middle of the USA, I had a long PSK-31 contact in Pittsburg PA. It works when your protocol is 31 baud (remember 300 baud modems? or 1200? Yea... that slow). Sharing 1 MHz over a thousand mile radius when the band is open would truly demonstrate some fascinating issues (and you wanna talk about problems with hidden nodes or implementing polling mechanisms over a 1000 mile radius?).
I'm mostly surprised that there are still believers in the dot-com model. Capacity gets paid for regardless of how you fudge the numbers. Either you buy the backbone capacity to feed it or you don't - and towers, trunking, engineering, licensing, compliance all cost real money. Ask a Level3 exec if they ever considered giving free dialup to "pay" for their national fiber rollout and watch their reaction.
This seems more like a frequency givaway scam in search of an engineering solution than anything. Remember that Internet CP80 Port stuff [deseretnews.com] that was so unworkable and absurd to anyone that actually understands anything about IP engineering? Our US house representative has indicated those people [cp80.org] are still pushing and they claim their engineers have looked it over and haven't had a credible complaint about their proposal. (Note: The CP80 people haven't gone away and have enough Congressmen that actually think their proposal is a good idea - good time for some followup!)
And if the discussions of the bandwidth utilization were correct, I'm curious where all that higher capacity stuff will come from that is necessary to revenue-share with the FCC. Or is that going to be an oops after the frequency gets allocated - guess we didn't have any left to sell and share. Thanks for the free frequency givaway for a single company. Now there's a scheme I can believe. Just send your Senator a few thousand dollars and get it earmarked. One of our own Senators was kind enough to find $50 million for a rainforest in the middle of flyover USA for a couple thousand dollar donation, so anything's really possible.
Re:Not enough bandwidth (Score:3, Interesting)
Exactly. Lots of folks don't realize we're in the bottom of the solar cycle. It's freaking dead on 10 meters (28 MHz) right now - I made a PSK-31 contact a month ago on 10m only because some other guy 40 miles from me was as curious as I was as to how dead it really was.
But the old hams talk about when the cycle wakes up. I mentioned in a previous post about working 17m (18 MHz) two nights ago for a 800+ mile contact. I picked up CW (morse code) from southern Mexico about 1300 miles south of me this weekend on 15m (21 MHz - just above the proposed frequency). I deal with wifi broadband engineering as part of my job and have enough to contend with in 802.11n/g/b/a, contention, collisions, etc. - spread my signal out and give it an occasional atmospheric skip and my network would be hell.
The point some of the "RF aware" are making is that taking a slice of 20 MHz isn't like taking a slice of 2400 MHz, or 20,000 MHz (which was what I originally thought they were talking about - great frequency for this, but requires *massive* cellular infrastructure - literally tens of billions of dollars to get decent national coverage).
Under 1,000 MHz, signals tend to start doing non-line-of-sight things, and depending upon how active the sun is, the lower you get, the more things get unpredictable.
Consider this: tonight we've got a storm system moving in. We were picking up solid traffic on 145 MHz (again, higher than the proposed 20 and less prone to these problems) from 300+ miles away. This is called tropospheric ducting - it has to do with warm air near the ground and frequencies travelling in a trapped manner within that warm air mass. This 20 MHz proposal would be amusing to watch it fail if it wasn't for the failure occuring during an abuse of the public commons.