Slashdot Mirror
Wi-Fi Times Sixteen
2Stupid2KnowIt writes "eWeek has a cool review of Xirrus' XS-3900 Wireless LAN Array. The unit consists of 16 Integrated Access Points and a wireless switch....all in one device. According to their website, Xirrus can achieve 800+ Mbps of bandwidth and handle 1000+ users. Finally enough bandwidth for us all to cut the cord?"
I'll never be able to cut the cord as long as latency for wireless is so high.
Parking lots will be overflowing with war drivers...
The NSA: The only part of the US government that actually listens.
Yeah right... until we can get 10-12 hours out of a laptop battery, we'll all have a cord. Might as well be a network cord with POE.
More
Sounds like a nice solution to save the pain in the butt of setting up large lan gaming events like the CPL... If a company can come up with a "gaming version" of this idea that "guarantees" lower latency and such i bet a lot of places will start adopting it... The only cords you'd have would be power cords... i like.
"Priced at $12,000, the XS-3900 is a relatively affordable solution for locations that require high-density networks. With all functions in a single device, administrators could see significant cost savings for deployment because multiple power and Ethernet outlets are not needed."
:/
So the savings on 11 ethernet jacks and power sockets are worth a $12,000 price tag?
"People" using "unnecessary" quotes should be "shot".
At that price, it's FAR cheaper to just buy the 16 devices and a router. But it does look cool.
;)
Not to mention you can microwave your coffee by just setting it on top of the thing
- Adam L. Beberg - The Cosm Project - http://www.mithral.com/
At that price, I'll prefer some Cisco Level 3 switch with 10/100/1000 speed anytime!!
No sig for now.
I hope people are careful where they use it -- using all the channels at the same time seems quite antisocial to any other networks that might be in the area.
For every problem, there is at least one solution that is simple, neat, and wrong.
Finally enough bandwidth for us all to cut the cord?
Multi-channel 802.11a has plenty of bandwidth to cut the cord. Even plain ol' 802.11g would suffice.
However, only one question really matters, and I doubt a positive answer:
Can it give me a decent signal more than one room away from the AP?
Turning on WEP, and authorizing on the MAC address would prevent most, if not all, wardrivers from connecting. Although MAC addresses can be spoofed relatively easily, it would definitely deter amateurish attempts at free internet access.
Hope be with ye,
Cyan
Sure they get a lot of bandwidth out of this thing, but don't they essentially get the same coverage they would with one AP?
For a large size area, the benefit to having multiple APs is that you can spread them out to increase your coverage.
>Finally enough bandwidth for us all to cut the cord?"
*Yells down to basement*
Kurt! You're moving OUT today!
You can't talk about Wikipedia's flaws on Wikipedia
Talk about bad math. :)
"People" using "unnecessary" quotes should be "shot".
Cutting the cord would be nice... This doesn't apply just to laptops people! Having one less cord coming out of the back of a desktop is great! Especially if you have a TON of desktops... Those messy cords add up!
Such gimmicky devices never take off. They proport to be some groundbreaking new amazingness when in actuality it's "a bunch of WiFi transceivers stacked on top of each other". That's not new, and it's not amazing. Companies have for years sold network cards that work by load-balancing traffic across multiple CAT5 lines - a good idea, sure, but it's never going to be widely accepted. How many double-100baseT NICs do you have? If you needed more than 100mbps, you'd buy gigabit ethernet. People who need more wireless speed are going to wait for the next step in technology, not a bunch of the same thing duct-taped together and put in a shiny plastic case.
Rex is 09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0
Is it just me, or is it ironic that war drivers would take over parking lots?
Hope be with ye,
Cyan
Am I the only EE (by training if not practice--I do software for a living) around that's a little concerned about the long-term effects of all this (additional--we've been absorbing UHF/VHF for 60+ years now) microwave radiation? Sure, sure, inverse-square law, skin effect, yadda yadda. I can't help but think we biological beings are much more sensitive to EMFs than the biologists assume. Could there possibly be a correlation between increased EMFs and the increase in autism, cancer, etc.? I haven't seen the actual research/figures, but I'm told that, when flourescent lights (with the older, dirtier ballasts) were introduced in the 50's, that learning disabilities skyrocketed (yeah, perhaps diagnostics just got better).
Nothing interesting to say...MUST...NOT...REPLY...ohtheheckwithit.
Okay, now this is going to sound incredibly stupid to the vast majority of you, it'll probably start a flame war and all of that good stuff.
The reason I won't use wireless is pretty simple, suppose I have my computer and my WAP sitting in the front room of my house. If you decide to pull up and park across the street you can sniff my data rather easily. Sure I can encrypt it, secure it, and slap an ACL on there so you can't get in or do anything with the data you capture, but the fact of the matter is you and your buddies hanging out in your car across the street from my house can sniff my data.
Now, if I've got copper inside. I pull up to the house one night and I notice the front window is open and there is some cat5 ran across my yard from your car window to my switch. I'm going to come out of the house, go to your car and proceed to knock the ever loving shit out of you in front of your friends. I'm not a big man, but if I was in that situation, I would be an angry one.
Of course, sure you can sniff my data with copper, but most likely you won't be doing it parked in front of my house, but rather at your own house which settles the whole notion of me dragging you out through your car window and kicking your ass there in the street.
Everybody knows that there is no such bandwidth as "enough"
What's the actual final performance for n=1 to 1000 users.
I'm 10' from an Airport (b+g), all bars on, lone user, and still usually just turn it off in favor of the Cat5 lying on my desk.
The 3Meg cable we use is underdriving everything - wired and wireless - but there's still a difference.
At a hotspot, checking my mail or surfing it isn't an issue - but at work connecting to servers etc. it's a palpable difference. Maybe we get spoiled like our ears with a stereo, but unless the actual experience matches cable, then keep the cable handy.
"Win treats sysadmins better than users. Mac treats users better than sysadmins. Linux treats everyone like sysadmins."
It's the reliability of the connection, stupid! Wired connections will ALWAYS be more reliable and more stable than wireless connections, period.
This is really a good idea. From TFA: ...permit only a single concurrent station to connect to each IAP...
... that might be a challenge. Like I said, similar, but not identical. Still a neat idea.
It's a wireless *switch*! Typical wireless deployments are like a hub-- 10, 20, 50 people connecting to the same AP. This is a really cool idea when you think about it. You're bridged solely to your own integrated access point, much like a port is your part on the bridge of a switch.
I say, get 12 WRT54G's at $60/piece, and a used/refurbed Cisco 2912, for about $200, load up the WRTs with OpenWRT, and you could probably do the a similar thing for about $1000. A little configuration and tweaking might be necessary though.
Also, don't know about the overlapping channels thing
FLR
>According to their website, Xirrus can achieve 800+ Mbps of bandwidth and handle 1000+ users
Since when is 11 equal to 1000?
It's way overpriced but i like this bit:
finally getting rid of insecure telnet!
This space is intentionally staring blankly at you
...that wasn't a dialup phone line.
Price to pay for living in a rural area. I'm surrounded by Amish so i doubt there is much of a CableTV/DSL demand out here.
do() || do_not();
Actually there are high end wireless routers out there that cost 3 to 5 thousand. They are intented for high security/authentication purposes. Also it generally cost 50-150 a drop for ethernet. And you are saving significantly more than 11 jacks as the intentions are for a thousand users. Easily making up the cost. Big question simply is, does it have enough distance for these thousand users, probably not.
Well, ignoring the fact that multiple clients can connect to the same AP, if you are in an office that doesn't have a wired infrastructure, or it needs upgrading*, then the cost of getting someone to rewire and patch could be quite high, especially if there are 100 people on the network (e.g., an office). 10000ft** of cat5e might not cost too much on its own, but the installation costs will cost a lot more, as would 8 decent 24-port GigE switches. And then all the laptop people will still moan about needing wireless access.
And think about having one of these at a convention or trade show... very handy.
* yes, I know that if the network 'needs' upgrading then there is probably a bandwidth need.
** yes, you can install switches at each block of desks to reduce cable runs. And so on.
No, it isn't.
I would expect such blatant racism on Fark, but on Slashdot? Mods please ban this asshole.
Seriously, why not just make it dynamic so you can just add more to your array and achieve more bandwith.
CS: It is all sink or swim...oh and did I mention there are sharks in that water?
Does it also function as a smoke detector?
Is it sufficiently tamper-resistant to not break when someone tries to check the battery?
cyn, free software and *nix operating systems enthusiast.
Not to mention the ~ $50 wireless card to get 54Mbps speed (to even be close to comparison to the 100Mbps of fast ethernet) per computer. Multiply by ~1000 users, and you have a starting ~ $62000 cost for network. And this is just the beginning.
Microsoft Sucks, F/OSS Rocks. I get mod points now right?
Parking lots will be overflowing with war drivers...
...and this is a problem how?
Zhrodague.net - I do projects and stuff too.
hahaha yes....Fark is the fountain of all racism indeed! Fuck You and Die
example.org - powered by Linux!
I don't understand. Scrolling through today's news I see three post dealing with increasing broadband access using extenders, power lines, and repeaters. But I still live in the sticks and still don't have any broadband access(except satelite option). All the big cities are the ones getting wifi first, bellsouth isn't going to spend all the money on dsl repeaters to put in rural area hick towns with only a few thousand people, and niether of the local isp's is willing to setup a wisp. They act like they care about the less fortunate(old, low quality lines) getting broadband in areas away from cities, but then end up keeping the technology for themselves.
Big question simply is, does it have enough distance for these thousand users, probably not.
Most fire codes won't allow 1000 users to use this AP.
Get your Unix fortune now!
Well, you don't need GigE switches. Wireless is just 54Mbps, and Fast Ethernet is at 200Mbps if full-duplex. A 24 port 100BaseT switches with 2 GigE port should be cheaper than 24 port GigE switches.
Of all channels, only 3 do not overlap.
How can they reach 800+ Mbps of bandwidth on 3 channels?
How is my company supposed to be able to afford the equivalent of a couple hundred T1s (ok, hyperbole) underneath this uber hot spot to handle all these users? Can't wait till one of our customers calls us this week and asks us to give him 800Mbits up and down. We already have to filter all spam perfectly without deleting a single one of their legitimate e-mails, and well as ensure they never get a virus or any spyware. Evar.
ah here's the link he "mesh'd" 3 or 4 open AP's to be one fat connection.
*shrug*
Build Your Own PVR/HTPC news, reviews, &
This thing is $12,000 - so it's obviously targeted as businesses. Now, imagine a cluster of employees tightly packed enough to justify 16 WAPs from one physical location. First off, the cheapest solution is to simply wire everyone up. 16 hubs, a switch and the required cabling and labor costs a lot less than $12,000. Even if they need wireless, why not just buy 16 wireless hubs at $60 a pop, and then hook those up to a switch? I don't get it why anyone would buy this...
I saw pretty much this exact same thing about six months ago. It was made by Vivato (a couple of Vivato people, being a local company, had come to one of my CS seminar classes). And it had been around for a couple of years by that time.
Not sure if they have an 802.11g version out yet. The one I saw (and touched!) was an 802.11b version with some nifty directional (phased?) antenna array stuff, using multiple Agere PHYs and an embedded PPC CPU running Linux to control the whole thing.
Several of the ones I saw (probably the VP 1200 or its predacessor http://www.vivato.net/prodtech_overview.html) are being used to power the WiFi network in downtown Spokane, WA.
Umm...802.11 specs only define 11 (for g, 14 for b if I'm reading my stuff correctly) channels. That means a couple of these radio's will be running on the same channels, which is about as pointless as you can get. Also running two APs on channel 4 and 5 next to each other will cause lots of interference with each other. So basically this thing will just be spewing lots of radio interference destorying the throughput on any single channel.
It's obvious this was not meant for a LAN / WAN indoor use. Who really would want to put up and foot the bill for a access point of this magnitude.
No one.
Companies could / might use it for inside their company...But being that it is alot cheaper to go with ethernet / wires.
Compare the costs of a huge switch / router with computers connected to it with a 12k AP that requires all of the computers inside the building to have a 50 dollar plus wifi card.
Not to mention the amount of interference / problems with wifi. Imagine the problems when their are close to 1k users inside a building.
Seriously, this solution isn't going to give you 800+ Mbps PER USER, each user gets at most what his/her wifi card will support (call it a max of 54 Mbps for g). The best that this can claim is "this is just as good as 16 wap's". As for 1,000 users, that's like having 60 or so people on a standard WAP.... love to see how that would work out.
Lets use an example with 100 users and the $8000 8-channel device (maybe the $4000 4-channel device would work well enough?).
... wireless is usually built into these devices, so you might only need 50 wireless cards for desktops. And if you've already got a wired network and just want to support the 50 managers with fancy laptops, then that's $400 per manager to save them looking for an ethernet jack and cable every time they have a meeting or whatever. Not really a big cost is it?
Ethernet:
Cable run average: $100 (from another post) = $10000
5 24-port switches: $500 - $2000 depending on quality
Wireless:
Device: $8000
100 Wireless Cards: $5000
More likely if you are deploying a wireless network solution you will have a lot of laptop users, wireless PDAs and phones too
The cost is working out fairly equal, to be honest. Well, the single point of failure issue might be raised, but at least it is easy to work out where the problem is!
As for having 10 separate APs ($1000), with 10 cable runs ($1000) and a single switch ($200) and installation ($1000), I'm sure that most managers would opt for the quicker solution that has less disruption.
no. that would be about 1000 cat5 cables. and say $5 a cat cable thats $5000 dollars.plus the pay for the techs to install all the cables. so i would say yeah its a good idea.
According to this web-site, using periods inside quotation marks in defiance of logic is an American thing. Canadians and Brits only put the periods (and similar punctuation) inside quotation marks where it makes sense. E.g., for the GP post the period makes sense outside of the quotation marks. However, consider this sentence:
You said, "They're in the queue, directly behind people who don't put periods inside quotation marks."
Here it makes sense to have the period inside the quotation marks.
It should also be pointed out that many Americans (myself included) follow the Canadian and British style.
Ben Hocking
Need a professional organizer?
"Nice gas you have here tho..."
--- What
What's the big deal about "cutting the cord"? Everyone's acting like this is the Holy Grail or something, but as near as I can tell, I'll still be sitting in front of the same old cubicle using the same old workstation, so what's so Evil(tm) about the "cord"?
Don't blame me, I didn't vote for either of them!
Network Security should be achieved by accounts, the entire key of which should be encrypted.
I believe (according to the manufactures) the typical range for 802.11a/b/g is about 100-150 ft. indoors (source). So how does this help people in a day to day work enviroment? I don't think many companies pack their workers that tightly together. For conventions and such it'll make setup a breeze, but you have to stay within the FCC regulations... so range would be an issue.
I say, get 12 WRT54G's at $60/piece, and a used/refurbed Cisco 2912, for about $200, load up the WRTs with OpenWRT, and you could probably do the a similar thing for about $1000. A little configuration and tweaking might be necessary though.
:-)
Yeah, a little configuration and approx. 20-50m (30-60yards) of cable
It's better to be the foot on the boot than the face on the pavement. ~~ tkx Kadin2048
Kind of like parking in your driveway?
At any rate I'm completely screwed. I ran kismet last week and detected around 60 wireless networks within range of my new york city apartment, that along with the cellular tranciever's on the roof of the building mean I'm blanketed 24/7. Good thing I already had kids.
music lover since 1969
I frequently put periods outside quotation marks on purpose, because I think that if the period isn't part of the quote (as in the quote is a complete sentence)it doesn't belong in the quotes. I know it isn't accepted officially but in order for something to be accepted officially, people have to already be using it that way.
You get 4 a/b/g APs, plus 12 a-only APs in their can. That's where they get the big # of users. For a total of 16 APs, this is a great device for high-density, conference room or public lobby applications. And the retail is more like $14K.
Oh yeah, telnet works on it unless disabled, too.
---- Teach Peace. It's Cheaper Than War.
Range of 100ft?
That covers an area of 31400 square feet. A typical office worker has a 10ft by 10ft cubicle, for 100 sq ft, so 314 workers could use this device.
I've noticed a tendancy for office space to utilise curvy tables and thinner desks with TFTs to cut the per user space down to around 10ftx8ft, or even 8ftx6ft. The latter would allow you to fit 620 workers within range.
TFA makes no reference to multiple (distributed) antennae - which was my initial thought on how range might be widened.
(But anyway, that would defeat the purpose of not having an ethernet system with multiple wireless access points, right? Certainly sounds less cost-effective than ethernet.)
Agree with parent - I'm also not seeing why this device has so much capacity when range (and therefore potential users) appears so limited.
For an example of EM that DNA does abosrb you need only look at UV radiation. DNA absorbs a roughly gaussian peak at around 260nm, falling to near 0 around 320nm. That's why short wavelength UV causes skin cancer while longer wavelength is OK (though it can cause opacity of the cornea). In fact, the method we use to determine the concentration and purity of DNA in a fluid is based on the absorbtion profile around 260nm.
Cody: You pay me $1000 a month. I live in my van in your driveway, eat your food, and at the end of the month I have $1000.
It's actually quite big compared to a normal AP but looks like nothing more than an oversized smoke detector when setup.
The term wireless switch may be misleading, more than one laptop can connect to each IAP and the wireless link is still a shared medium.
Where this product differs is it's ability to use all of the unlicensed spectrum within a given area. This translates into 3 channels on 802.11b/g and 12 channels on 802.11a. The range for 802.11b/g was about average but for 802.11a it was great. This is because each IAP has it's own antenna pointing in a specific direction (70 degrees wide) which allows the signal to be amplified by 7dbi rather than the normal 2.2dbi for an omnidirectional. This translates in the transmitted power being roughly doubled.
Some people say they could "emulate" the devices result using a bunch of WRT54Gs but since those operate in 2.4GHz they would all overlap and cause massive interference problems. The only effective way to get massive amounts of wireless bandwidth within a given area is using 5GHz because there are more non-overlapping channels (12 vs 3).
Another cool feature with the product is the lights on the front of the array. There is one for each IAP and they light up when someone associates to that IAP.
Not only would you be obliterating all other wireless networks within range (this thing is active on all the channels) and getting collisions slowing it, but there is no way in hell you are going to be getting the advertised 54 Mbps on each channel.
Now I have never used the 54 Mbps stuff but from my experience with the 11 and 22 Mbps equipment I can say that you get no where near that speed even with the antennas nearly touching each other. My 22 Mbps network gets around 6 Mbps in actual use, I have never seen it go above that. And when I was on 11 Mbps it literally topped out at exactly half that speed (3 Mbps). It seems like we have all been duped.
The ratio of people to cake is too big
So, you put this on the 2nd floor of your building. It'll be in range of the almost 314 people on the 1st floor, and 314 people on the 3rd floor.
Eating a whole bottle of hot sauce every day.
Exactly.
Also, I've just read another review, and the 802.11a reception was 22mbit at 100ft from the AP, and it looked pretty good at 150ft. So even if you were in a normal office building (instead of a 100ft radius cylinder, heh) it would reach pretty much everyone on at least 2 floors. For an L shaped office building, 200ft long by 80ft wide (floor area = 26500sq ft, which is 265 - 500 employees per floor)) it would be ideal
Could you wrap a pringles can, or such, around each antenna, so that each covers a different direction, thereby increasing the range?
Actually there are high end wireless routers out there that cost 3 to 5 thousand. They are intented for high security/authentication purposes. Also it generally cost 50-150 a drop for ethernet. And you are saving significantly more than 11 jacks as the intentions are for a thousand users. Easily making up the cost. Big question simply is, does it have enough distance for these thousand users, probably not.
With this kind of bandwidth it becomes feasable to set up (relativly cheap!) repeaters near the edges of the main connections range, each repeater would naturally be running at much slower speed, but you shove a dozen users on a 802.11g link, even assuming you are only getting 4MB/s over it, that is still over 300KB/s per user!
Need help treating your acne? Come here!
Hmmm, Wifi times 16...
According to my math, 802.11 x 16 = 12,833.76
I guess we get discounted down to $12K for buying in volume?
Now I have never used the 54 Mbps stuff but from my experience with the 11 and 22 Mbps equipment I can say that you get no where near that speed even with the antennas nearly touching each other. My 22 Mbps network gets around 6 Mbps in actual use, I have never seen it go above that. And when I was on 11 Mbps it literally topped out at exactly half that speed (3 Mbps). It seems like we have all been duped.
I'd have to disagree. I have a 52 Mbps 11b/g wireless built in to my AMD laptop (eMachines) and I frequently get 33 Mbps or 52 Mbps, and that's piggybacking off of my neighbors or the coffee shops down the street.
Just bought a 108 Mbps wireless box with 4 hard LAN ports out so my old PCs and my new laptop can share it and I'll see what it actually cranks out, but I'm not too concerned.
-- Tigger warning: This post may contain tiggers! --
Or driving on the parkway.
excellent!
I think the parent was saying 11 jacks because this unit is several wireless AP's combined into one unit. What does this unit provide you that plugging in 12 separate and far cheaper access points could not? (although the blurb claims it is 16 access points, not 12, but either way...) afaik, you can buy an 11g access point for around $50 these days, can't you?
If I don't put anything here, will anyone recognize me anymore?
"Priced at $12,000" Woah, that is bullshit. They're only charging so much because they know they can get away with it. They could probably easily sell the thing for $100 or $200 and still make a profit. At such a price as $12,000, only businesses will be able to afford it!
That just means thats increased revenue for those who are developing products to cure cancer, autism, etc. You're creating new jobs and creating revenue! Or so my capitalist overlords tell me.
The Doormat
If you're not outraged, then you're not paying attention.
TOO LATE
music lover since 1969
Some of the 11Mbit/54Mbit bandwidth is used to send the data (headers,error correction,whatever). So, your applications cannot use the entirety of this space.
That would be 12833.76!
(802.11 x 16) just to show my work.
Where does the school board find them and why do they keep sending them to ME?
first it's 54mbps. second just because the little icon says "Connected at XXmbps" doesn't mean you're getting that in actual throughput. And you've certainly got no way of testing that with a borrowed internet connection (that is no doubt much slower than 54mbps). on top of that 802.11g doesn't ever get close to sustained 54mbps, more like 20mbps in optimal situations.
First of all, WiFi does not have 16 non-interferenced a/b/g channels allocated now.
Secondly, each channnel has less than 25Mbps throughput.
So in summit, this device will delieve less than 400Mbps traffic at it best situation.
first it's 54mbps. second just because the little icon says "Connected at XXmbps" doesn't mean you're getting that in actual throughput. And you've certainly got no way of testing that with a borrowed internet connection (that is no doubt much slower than 54mbps). on top of that 802.11g doesn't ever get close to sustained 54mbps, more like 20mbps in optimal situations.
I don't know, i kind of thought massive downloads of WinXP patches kind of worked pretty well for benchmarks, since I did that after I firewalled the laptop over the wireless connection, and I knew how big they were.
But, yes, as I said, it tends to vary between about 22 and 33 MBps and sometimes 52 MBps. Only occassionally do I get 11 Mbps.
-- Tigger warning: This post may contain tiggers! --
[turn on WEP and MAC => almost safe]
Depending on WEP is a joke, depending on MAC addresses to limit access is even a bigger joke. So why aren't you modded funny?
so you still do not see the difference between your connection "speed" and the actual throughput?
A benchmark would be going to a site that provides you with a bandwidth report of some kind.
Besides all of that if you use a borrowed connection that is say 10mbps (most DSL in north america is 1.5-8mbps) you can't possible make a claim that you get between 22 and 33mbps.
For the record a "52mbps" (802.11g is 54mbps) connection would
work out to 6347.65625 KILOBYTES per second or about 6.12 MEGABYTES per SECOND. this would me a 60 megabyte download would be done in less then 10 seconds.
Knowledge is power.
you need more power.
i didn't say i borrowed the connections (and there's a bunch) when they were using them.
besides, as I said, i just got a 108 Mbps wireless hub, so we'll see how it handles my 8 Gbps cable modem splitting, both single use (with the other Macs and PC boxen off) and split (with the Mac and PC boxes on).
I also have a DSL line, and that's slower than my wireless is.
It could be because I live in a valley called Fremont in Seattle, where wireless is cheap, strong, and organic fair-trade, like our coffee. Maybe we have less interference due to our geography and all the nice high-test wireless provided - we also have Internet 2 for some of the UW offices in our neighborhood which is even faster.
As I said, your mileage may vary, I don't live where you do, and all I know is relative speed comparisons at different signal strengths and in comparison with my cable modem and DSL speeds.
-- Tigger warning: This post may contain tiggers! --
I read they have more models than the $12,000 beast. The 8 radio version ($7,000) would be great for our campus + they're giving free evals. Rather than hypothesize about the product -- mine's on it's way. I'll let you know how it goes.
I just happen to have finished building an FPGA based system which can handle multiple WiFi channels. (5 channels per board, number of boards only limited by how many PCI slots available.) Anyone interested in it?
So, 'Metal Health' really will drive you mad!
It is most certainly "officially accepted" to put the punctuation outside of the quotes when appropriate grammatically. See other posts in the grandparent's subtree for links and whatnot. If you get marked down on a paper or something for incorrect quoting style, throw some "colours" in there and tell them you're kicking it UK-style.
My 54Mbps wireless usually get about 33... I have *never* seen it go to 54Mbps. My wired network (100Mbps) has transfer speeds of a little less than 10,000KB/s, which I figure to be about 80% of the rated speed. My wireless connection (54Mbps) gets transfer speeds around 3,000KB/s, which is closer to 45% of the rated speed. You are either very lucky, or wrong, with your wireless reaching it's rated speed. Enjoy your 8 Gbps cable modem...
My 54Mbps wireless usually get about 33... I have *never* seen it go to 54Mbps. My wired network (100Mbps) has transfer speeds of a little less than 10,000KB/s, which I figure to be about 80% of the rated speed. My wireless connection (54Mbps) gets transfer speeds around 3,000KB/s, which is closer to 45% of the rated speed. You are either very lucky, or wrong, with your wireless reaching it's rated speed. Enjoy your 8 Gbps cable modem...
well, as I said, it could be where I live, we're one of the most wired and unwired neighborhoods in all of Seattle, where Getty Images and Adobe are located, and have more Thai food and free wireless hotspots than almost everyone. Fremont, in Seattle.
Here at the UW we have wide fat pipes that make cable modem look sick. Especially at the UW Medical Center, cause we process genomes and do long-distance surgery and that kind of thing.
-- Tigger warning: This post may contain tiggers! --
"Finally enough bandwidth for us all to cut the cord?"
m ain) and breach every wifi system on the market I can think of, all it takes is time.
Your a tech of some form I am assuming since your writing a slashdot article, and yet your think bandwidth is the reason large chunks of us are staying away from Wifi. Ahem. Interesting theory.
Try this instead, if I have a cat5 running from my PC to my router I can see the cord.
I know if theres some vampire tap going on. Someone has to physically break into my home (which I can add multiple additional layers of security too to not risk a single point of failure of a dodgy old doorlock) and connect something to it so I have a damn good chance of spotting something that is intercepting my bandwidth.
Wifi, unless I live in a faraday cage I am sending out signals that say "Oh please come and take a look at my network and start using it" to everyone in a short radius from my house or anyone going past who is into wardriving. It then becomes a matter of not IF someone is able to break in but a calculation as to how long it takes even with WEP and MAC addressing alot systems can be breached in 10minutes to 2 hours if they are high traffic. Any old script kiddy can go to Knoppix STD (http://www.knoppix-std.org/) or Remote Auditor (http://new.remote-exploit.org/index.php/Auditor_
Damn, why don't we just network our homes and offices with that stuff?
Couldn't agree more!
Today's 'desktop replacement' models might pull 2h on a new battery. Today's mobile systems might pull 3.5h. This is just doing 'word', but you might get half that playing a video game or watching a movie. Double the batteries (at $175), weight, and replacement cost-- and you might get more.
Not to mention a performance hit (lowering clock speed, CD spin speed, etc) when running on battery.
Still impractical. Even when I work outside I leave the laptop plugged in. A plane is useful playing a game or DVD for 2h.
I'd sacrifice speed for power in a mobile unit case, but I'd love both. Nonetheless, wireless is a joke. It's got so many problems (especially in a large office or larger house), inconsistancy, serious latency, and then of course power troubles.
Find if you _NEED_ wireless, but for the few hundred to wire every room in your house with CAT-5e and some wall plates/crimpers, wireless is a joke.
-M
The 802.11A propigation would drop in the next room but the 802.11G would carry on, keep in mind that with a full power Cisco AP and a full power Cisco card you can go about 90 feet and still achieve 54 MBPS (actually half duplex, more like 21 full with WEP). If you are using AP's that are not Cisco, Symbol or Proxim your signal is not going to penitrate all of the walls.
When I do a wireless network, if the user density is high I will use 802.11A since the radio is so weak, I can get less users per cell or AP. If I am seting up a network for maximum coverage I use 802.11B/G to get maximum propigation.
Use 802.11A in a Cube farm, 802.11B/G in a warehouse where it carrys far and only has to communicate with 5-10 scanners or laptops per AP.
ack! I've been had by a troll!
Symbol has had a similar device that handles 48 access ports (access points stripped of their switching hardware) for about two years now. That's three times what this does.
Yes, it's all in one device. To me, that's a bad idea. With the Xirrus device, you're forced to have all the transceivers in one place, even if it would be better (due to shadowing and multipath issues) to split them into two groups. With the Symbol device, you can split up the access ports however you need them, to cover dead zones, etc.
I mean, you did do a site survey before installing your large-scale wireless network, didn't you?
(Disclaimer: I used to work for Symbol, but I'm not really a Symbol promoter. I don't even have stock.)
... to use a single device to deploy a WLAN using all nonoverlapping channels in the 5GHz (IEEE 802.11a) and 2.4GHz (802.11b/g) spectrums ...
Most of the devices in 2.4GHz (Bluetooth, Zigbee, WirelessUSB(developed at Cypress Semiconductor) implement some form of channel sharing. With Xirrus's IAP (integrated APs) all the channels could be occupied and this would imply that there is no free channel available for other wireless devices. For example, you would probably see a huge latency using your Bluetooth mice in such an enviroment or for that matter using your Bluetooth enabled PDA!
as computer theory grammars. No, I agree - I prefer your method as well. But I'm not going to push my luck by using that any papers I submit to journals, etc.
Ben Hocking
Need a professional organizer?
Official range specs for wireless networks is 100m, or about 300 feet. This device uses 16 directional antennas to produce what is effectively an omnidirectional source (Each antenna covers 1/16 of the circle). The company claims that extends the range by 4x.
That is a theoretical radius of 1200 feet, not 100 feet. Do the math.
There were two links, one was to the official site. Each of the 16 antennas are pointed in different directions, and are unidirectional. They only simulate an omnidirectional source because each antenna covers a slice of the circle. The company claims that because each antenna is unidirectional, they get 4x the range.
They don't actually say if there are 16 antennas, only that there are multiple unidirectional antennas in different directions. 16 access points, so 16 antennas would make sense, though it's not certain.
I can see a use for this kind of density is in extremely crouded and busy environments similar to a trading floor at a stock exchange or something like that. I would be willing to bet that one could use multiple of these devices by simply rotating them in relation to eachother so that different channels over lap: C1 -- (AP) -- C12 C1 -- (AP) -- C12 In this situation C1 and C12 would overlap and thus not interfere... the other channels being directional would not occupy the same space and interfere with each other.
Sometimes the best solution is to stop wasting time looking for an easy solution.
not to forget the fact that in a wired network, the wires can fail. 1 idiot contractor with a drill or a few rats or mice and the wiring would need to be rerun. Also consider that case that some buildings being brought up to having a wired network might not have ideal wall cavities for it.
Humans are slow, innaccurate, and brilliant; computers are fast, acurrate, and dumb; together they are unbeatable
How does this not violate the FCC restrictions on power? Granted you could buy 2 routers and put a peice of metal in between them and say the same thing but this is sold as a single package.
--
WHO ATE MY BREAKFAST PANTS?
The minute I saw this article, Vivato was also the first thing to spring to my mind.
:-) I think I remember the sales guy mentioning to us that one of their customers managed to port Kismet over, even.
We had one of the Vivato sales people visit us last spring. (We're also not too far away from Spokane ourselves.) We couldn't justify the cost of the system at the time for the application we were thinking about using it in, but we got to see and play with possibly one of the coolest 802.11 devices in existence.
BTW, the Vivato system (at least their 11b one) uses 14 radios...13 to talk back-and-forth to clients, and one for passive listening/"rogue AP detection." It is built around Lucent/Agere Hermes radis, has a PowerPC for the main processor, and runs on Linux.
It seems to me that Vivato's phased-array antenna system is probably at least as good, if not better, than the Xirrus's uber-sectorized solution. Their claim of 16 non-overlapping channels seems amazing at first (are the only considered non-overlapping because despite high channel reuse the sectors don't cross each other's paths?), but as far as I can tell, they're counting the 5GHz spectrum in with those non-overlapping channels, which means two things: A) you HAVE to have a dual-band card in your laptop if you want to take maximum advantage of this thing, and B) there's naturally going to be some inconsistency in performance since 2.4GHz and 5.(2/3/4/7/8)GHz react differently to physical obstructions and have different rules set by the FCC for maximum EIRP and such that both the AP and client have to honor.
Vivato...*drool*. I want one for my home.
-- Nathan
Well affordable my hairy butt!
Java Oracle Linux Enthusiast
Thank you for the figures.
So one of these $12000 devices can cover 4.5m square foot, or 1/6th of a square mile. A normal AP can cover 1/100th of a square mile.
Yes, you could use 16 APs to cover the same area, with 16 cable runs, 16x as many people disrupted, 16x as many locations to administer. 16 $100 (decent) APs with 16 $100 cable runs and a decent switch. That's probably around $4000 before ongoing running costs (33x as many things to go wrong).
And consider your typical metropolitan wireless network. It is unlikely that you'll find a location every 1/100th of a square mile to install an AP, but you'll find a location every 1/6th of a square mile. Installation costs per AP are probably a lot higher too, what with needing a network to connect to at each location. Even in a residential estate (one house per 5000 square foot say, you can encompass 900 properties. Not bad if one in five are each paying you $10 a month for access - you'll break even in a year. In the UK where a residential estate typically has plots of around 1500 square foot you'll cover 3000 properties - if you got one in 10 paying for wireless service you'd make back AP costs in 4 months, and hopefully the $3000 a month after that would be more than enough to pay off upstream bandwidth costs. Run a few of these networks and you'd have a nice income. Great for smaller towns that don't have broadband access too.
Good points. I remember when I worked at BT that the building I was in which was built in the 60s had to have a false floor put down during refurbishment so as to allow easier laying of ethernet and power cables (they were changing to open-plan from walled laboratories).
It's You're not Your.
ack! I've been had by a troll!
dude, i'm not arguing with your experience with wifi speeds, i'm just attesting to my experience with wifi speeds in the b/g range.
-- Tigger warning: This post may contain tiggers! --
You misunderstand. My figures were for 16 access points all in one spot. You only need one cable run to get to the cluster, from there you have the 15 cheap 3 foot (or shorter) cables. Don't forget, each of these $100 access points are also 4-port switches.
One thing I might have overlooked is the cost of 16 unidirectional antennas, then you'd have the exact equivalent, hardware wise, to this $12000 solution. Maybe toss in the money for an enclosure of some kind for the 16 access points.
About the only advantage the $12000 solution has is that it is probably less work to configure. It is up to you if that justifies the 6x cost from the do-it-yourself solution that is equivalent virtually all other respects.
I don't pretend to know the FCC regulations, especially because I'm not in the US. But I'd suspect that, for one thing, the power limits are per frequency. This device uses 16 different frequencies, 8 on the 2.4ghz section and 8 on 5ghz.
Also keep in mind that home routers like those from Linksys are WAY under the FCC limits. You could probably pump up the juice by several times before you hit the FCC limits.
You mean warparkers, right?
Karma: It's all a bunch of tree-huggin' hippy crap!
And volume. In the UK BT installs its OpenZone wifi access in the top of its extensive network of telephone boxes. There isn't much room there.
And you still have the problem of having 16 devices to troubleshoot, 16 antennas to aim, 16 cables to ensure are working fine, and a switch.
But yes, if you have the area for the setup, and the time to configure it, that solution is fine.
But if it's like every other wireless device, you'll see 40% of that at most. I'd be happy if my 54Mbps wireless ran at 75% of it's theoretical speed.
This is true, however it is difficult to explain where the other $10,000 dollars in the price comes from. You would expect that to take the parts of 16 access points and put (some) of them onto one large PCB would cost a lot less than 16 seperate devices with all the redundant hardware that entails.
Of course, does BT really need to support 1000 people per phone box?
That last bit isn't true at all. The FCC limits are stricter than you think. Many of these routers have firmware patches to increase the range--do you think that if that was all it took linksys wouldn't do it themselves? They don't because it goes well over the limits when you even double the power (let alone increase it by "several times").
--
WHO ATE MY BREAKFAST PANTS?
It is probably an economy of scale thing.
Then again, if these things cost half the price they are being sold at, I reckon they'd sell 4 or 5 times as many.
BT might not need to support 1000 people, but in some locations it might be their best option for providing broadband connectivity to a rural community. The $4000 unit could be ideal for the one phone box remaining in the community and support up to 250 users.