Slashdot Mirror
5 GHz Wireless Networking With CMOS Transceivers
cthugha writes: "On the back of IPv6 and fat pipes, we Aussies have been at it again. Radiata, a company set up by a couple of Sydney-based researchers, has achieved wireless networking for LANs in the 5 GHz band using CMOS-based transceivers. This means (i) low power consumption, (ii) high bandwidth (currently, 54 Mbps with a view to getting up to 100 Mbps) and (iii) low cost. Unfortunately, like most Australian inventions, this one has only found serious commercial backing overseas, specifically from Cisco (government/big business over here has no brain)." Products, please? For half a billion dollars' investment, I hope Cisco plans to start selling some toys, fast.
Radiata Product Briefs
At 5GHz, and certainly with low power, they'll probably be local to an area of a few hundred feet. But it also depends on how wide a spectum they need. (I think the FCC is auctioning a band near 5GHz.) I'm getting my guess of a few hundred feet from assuming they'll use frequency hopping spread spectrum techniques similar as to what's available now.
For a look at the US spectum allocation, as it was in 1999 at least, check this.
US FCC Spectrum allocation as of 1999.
Still 54Mb at duplex isn't bad. 11Mb or 44Mb and somtimes 54Mb?
But since they use (2) bands at 20MHz they're gonna need a Jeff Bezos ego sized chunk of spectrum to make it fly. To make it worse for both Cisco and their young(er) charge every damn countries spectrum is chopped up seperately, and they're getting kinda crowded. Given that the spectrum these puppies go at is an, as of yet, unallocated and (to my mind) a nessecarily large, block, I've gotta say these are at least 2 or more years off from the consumer.
There are a couple of up sides, these toys are really low power, they'll have very small antenna, and by the time you can get one you'll be playing 3d interactive web games off your G3 phone and won't give a crap.
--Jimmy has fancy plans; and pants to match.
I would've tried the Linux WLAN support with my older ZoomAir 802.11 card (Harris Prism chipset) but I can never connect to the ftp site where the drivers are to download them. Actually, I guess now they have it working via http so I'll have to give that a shot at home. Maybe then I can replace Win98 on my adhoc based 802.11 access point I built painfully out of a P5-90 with 16 megs of ram with Linux. Though, oddly enough I get decent performance out of Win98 as an accesspoint if you turn ip forwarding on and don't have the box doing ANYTHING else. :-)
IIRC, while parts of the 5 GHz band are still being auctioned off, there is already a portion of the spectrum that is reserved for unlicensed communications of the same category as 2.4 GHz WLANs/phones, etc. (The regs are linked to from one of the WLAN HOWTOS, I believe, can't remember exactly where.)
retrorocket.o not found, launch anyway?
The reason your lan doesn't boil your brains is elementary.
802.11b network card: 0.1W (100mW), open air. Power decreases as inverse square of distance from transmitter.
Microwave Oven: 600W. Contained in a microwave reflecting chamber so all radiation is absorbed by molecules in the food.
I'm running Win95A with DUN1.3 on a 386SX/33 with 4MB RAM and a 120MB HD as an access point. Doesn't even take that long to boot--it's amazing how lean Win95A was by today's standards. And it's even more amazing how few people know about its secret routing capabilities; I've never met anyone in the flesh that knew, only on the net. It works with 4MB because it doesn't really do much while routing; once you actually try to bring up apps and use them, you quickly notice what hardware you're on.
Erm... just to clarify a few points in your post...
UHF or VHF were just the frequencies that the streaming content was transmitted upon. The actual protocol used was a rather primitive compressed format termed NTSC (short for Never Twice the Same Colour, in reference to it's complete lack of any mechanism for correcting common colour errors caused by phase distortion in the transmitted signal). While the low latency and compression of the format were technically superior, NTSC has basically no error-correction facilities, an essential part of modern streaming formats.
Side note: NTSC and it's competitor PAL had the capability to transmit mono and colour content in the same signal. Not bad, eh!
US Frequency Allocation Chart
That has a much clearer picture of what's up in the 5ghz range - Aircraft Navigation being the primary use, with bits of Maratime Nav and a little slice of "Amateur" in there near 5.8ghz.
I'd be most worried about the line of sight issue. The higher your frequency, you can send more data, but more things are going to block your signal. Something operating at this frequency is going to have problems going through a wall or a pane of glass. It might work ok through a wood frame/drywall construction house's interior walls, but I suspect that if you've got steel frame or even lath & plaster with hardware cloth backing, you're not going to get very far, or get very good performance if you do.
Firstly, at 2.4GHz there is a lot of stuff as the band is unlicensed. Microwave ovens emit in the same band. So there is a high probability of interference with any of the 2.4GHz products, that will only get worse as BlueTooth starts to be accepted. In comparison the 5GHz bands are licensed for use with WLAN products, with similar bands licensed worldwide. This basically means that the 5GHz products have to play fair with each other.
Secondly, there is a lot of bandwidth available at 5GHz, especially in the US and Europe (Japanese have a lot less), for WLAN applications. Given that the amount of data you can carry is pretty much proportional to your bandwidth (ignoring noise for the moment) this gives potentially enormous resources. For instance the American natioanl infrastructure bands are 5.15GHz to 5.35GHz and 5.725 to 5.825GHz, giving a total of 12x20MHz channels each capable of 54Mb/s. In Europe you have 19x20MHz bands. All of this bandwidth is liscensed NOW worldwide for this style of product.
Then you have to consider the type of coding used on these 5GHz systems. They are all OFDM.. Effectively, this means that those 20MHz bands are split up in 64 seperate channel (some used for guardbands) and each of the 64 channels is modulated seperately. So all those horrible indoor propagation effects that cause channel fading are less importance. In a traditional single carrier system multipath can kill the channel, but with OFDM you lose a couple of channels which are then recovered with error correction
All in all, I'd prefer one of these 5GHz systems over anything in the 2.4GHz band.
D.
They forgot to mention (iv)Lack of ability to penetrate through objects.
As you get higher in RF bandwidth, you lose the ability for the RF waves to penetrate objects (walls, furniture, etc). So, while we will (theoretically) get higher bandwidth, we will also sacrifice distance when compared to similar power consumption levels over lower-bandwidth technologies.
-This sig intentionally left blank
This wireless networking just gets better and better... I still have a cell modem i my laptop, and it's truly painful to use, especially sitting at my school (UU) library, which has wireless ethernet. But for many of us /.ers, the deciding factor with these new technologies is whether they become standardized. 802.11 was the greatest thing that happened to wireless ethernet; if we really want an idea like this to take off, it has to be standardized and widespread from the minute they get the first product out the door. With that said, let's start seeing those products!
I've had this sig for three days.
My parents had a TV in the 60's that worked on similar technology... apparently the image and sound were streamed to the this video-on-demand appliance using a protocol called UHF (or it's competitor VHF). In the same vein, I must note the my grandfather in fact built his own wireless audio streaming appliance back in the 30's! In fact, the appliance was called a "wireless" and was based on an Amplitude Modulation protocol called, simply, AM.
Isn't progress amazing?
2 1337 4 u!
If the Australian Government did Invest in the Technology, Cisco would then buy Austrialia.
:^>
-
--
A U.S. company (see www.atheros.com) demonstrated working 802.11a radios at the most recent N+I show in Atlanta (Radiata chips were not showable at that time). Both companies use "standard" digital CMOS to create a two-chip 5 Ghz radio. One of the chips is all-digital and provides the DSP and OFDM modulation functions at baseband frequencies. The second chip is the 5-Ghz analog transceiver - which is the trickier bit. The main differences between these two chip sets are first: the RF power amp is incorporated into the Atheros design - it must be an external component in the Radiata design. For indoor use in the lower 5 Ghz bands this is a 50mw device. Doing this in cmos is even trickier than the transceiver design. Secondly, the Atheros pair of chips incorporate a full NIC/MAC design into the digital chip, meaning that it has pins to hook directly onto a PCI bus or a Cardbus. The Radiata chips must be accompanied by an external MAC/bus/dma device. The Radiata digital chip will interface to a standard Intersil MAC chip, which at this time, is not able to drive the radio at full rated bandwidth (54 Mhz). The Atheros cards have been demonstrated at 72 Mhz (using two radio channels in parallel rather than just one), and are expected to achive higher bandwidths (108Mhz) in production. Be wary of the difference between channel bandwidth and data troughput for this class of network. With ethernet and other wired LANs we're using to achieving data throughputs near the line rate. With radio protocols there is a great deal more wait time in the protocol. Current designs do well to achieve 50% of the channel. Note also that all these devices are half-duplex since it seems to be impractical to make a transceiver that can receive while transmitting (on the same channel). In the future it may very well be standard to build a chip that contains several radios operating on different channels in order to get both more bandwidth and full-duplex operation. But for now, it is not practical. There are several frequency bands at 5Ghz set aside for non-licensed operation. The low band specifies 50 mw power at most and is intended for indoor use. Higher power levels may be used in the upper bands - up to one watt in some cases. Because of its dependence on an external power amp, the Radiata chips are well-suited for use in point-to-point external links: imagine you're Cisco and you'd like to provide wireless links between routers. In contrast, the Atheros design is solidly focused on the low-power world of laptops and home/office applicatins. In the near future one might expect a next-gen Radiata design to resemble the Atheros design, and a next-gen Atheros design to have some of the properties of the Radiata design. Range: the range of these radios at 50mw indoors is dependent on many factors. They seem to be more robust than 802.11b - because of OFDM they are much less susceptible to multi-path interference. On the other hand the penetration through construction materials at 5 Ghz is much less than at 2.4 Ghz. Range seems to be 100-150 feet indoors - mileage will certainly vary. Signals seem to reach their destination by reflecting off walls rather than by penetration. Summary: the Radiata chip set is a two-chip radio. The Atheros design is a two-chip NIC card that incorporates a 5-Ghz radio: just add antenna and power. Taken as a whole, these two designs represent the state-of-the-art in integrated signal processing and radio design.
Microwave ovens operate at 2.45 Ghz.
This wavelength is, contraty to popular belief, NOT the 'resonant frequency of water molecules', though in vacuo, an h2o molecule does have one resonant frequency near this (if memory servers). In liquid form, this gets far more complicated, and isn't useful. Also, if this WERE true, microwaves would not penetrate the object, but would be stopped by the first layer of resonanting molecules!
At this frequency, polarized molecules (such as water) are shaken as the field oscillates. And shaking molecules = heat, right?
The fact that microwaves operate at 2.45 Ghz is, I believe, the reason that the 2.4 Ghz ISM band exists in the first place. The band that is now used for a lot of htings, including wireless LAN stuff.
I've noticed a bunch of discussion on this, so I thought I'd dump a bit of reality on to the situation... First of all, there are currently two vendors that have working alpha chip implementations. The first is Radiata, which is discussed at length in this thread. The other is a company called Atheros (http://www.atheros.com). Both vendors create two-chip wireless CMOS solutions to operate in the 5 Ghz Unlicensed National Information Infrastructure band (UNII). Both vendors are implementing what is known as the IEEE 802.11a standard for wireless networking. This standard was ratified at the same time as IEEE 802.11b (June of 99 or so). However, unlike 802.11b, which sends 11 million bits per second through the air using a flavor of direct sequence modulation, 802.11a uses OFDM ((coded)orthogonal frequency division multiplexing)... they drop the C from the acronym for asthetic reasons. To answer some of the questions that have cropped up: 1) 802.11b and 802.11a use the exact same MAC. For the non-network literate, that simply means that the frame headers are arranged in the same way, and both use CSMA/CA (Carrier sense, multiple access with collision avoidance) as their method of transmission at layer-2. However, since 802.11b operates in the 2.4 Ghz spectrum, and 802.11a operates in the 5 Ghz spectrum, there is no chance for interoperability. The good news is that if someone ever writes a good wireless sniffer (ala TCPdump, but for the MAC layer), it will work with both .11a and .11b.
2) In terms of distance, the 5 Ghz band is broken up in to three subsegments. At the lower end, there are two 100 Mhz 'sections'. The first section at 5.150-5.250 Ghz is limited to 50 milliwatts EIRP (Effective Isotropically Radiated Power). The second section, at 5.250-5.350 Ghz, is limited to 250 mW EIRP. The third section is way up at the top, from 5.725-5.825 Ghz, and allows up to 1 W EIRP. The products from radiata and atheros operate in the lower two sections, with a maximum power of 50 mW across the whole 200 Mhz. There are several reasons for this, mostly due to the fact that for proper channel selection, it would be difficult to move from one channel to another if you had to change power output... In any case, 50 mW on one of these radios is enough to get you roughly 300 feet.
3) It is important to note that while these radios
will offer a maximum distance of about 300 feet,
you won't get 54 Mbps at that distance (54 Mbps is the highest speed that these radios implement on a single radio channel). The 802.11a standard specifies a number of encoding and modulation techniques which result in different bandwidths from 6 Mbps, to 12, 18, and 24 Mbps. Since the amount of frequency being used isn't changing, something else must be. In this case, it is the number of bits being transmitted per oscillation. (1,2,3, or 4). There is a direct trade-off between signal complexity and distance. The more complex a signal (the more bits/oscillation), the more quickly it degrades, and the more succeptible it is to interference. Thus, at 100 feet, you might get 54 Mbps, but at 120 feet, the signal becomes degraded, and so the radio drops to a less complex modulation scheme, and your throughput drops to 24 Mbps.
Hope that clears up the issues regarding distance. Incidentally, the high 100 Mhz are 'reserved' for point to point wireless applications (at 1 watt). These have the potential to go up to 2 to 3 miles or more, depending on antenna design.
Another interesting thing to note here is that while Radiata is an australian based company, they had planned on targeting the US market from day one... 802.11a is an IEEE standard, and is currently only going to be adopted in the US.
In contrast, HIPERLAN/2 is the european standard
of choice. It does NOT share a MAC with 802.11b, which is one reason the HIPERLAN/2 system is still under development. HIPERLAN/2 works on a time-division multiple access system (TDMA), and has its roots in "wireless ATM". (ATM being asynchronous transfer mode, not cash machine protocol). HIPERLAN/2 offers the promise of better guarantees for signal (QOS), because there is no contention for channel amongst devices.
Let me keep rambling for one second more... Recently, Intelsil (the former radio electronics group from Harris Semiconductor) announced their own IEEE 802.11a wireless roadmap. The moral of the story here: The incumbants in this segment of the industry (intersil, lucent) are quietly making their own plans. 802.11a will take off, it will be big, it's just a matter of when.
within 10 feet
of a hub/repeater/etc (perhaps including another end unit). (This assumption rises from the fact that they're obviously very proud of the capabilities it
So, with that in mind, a list of the top ten activities you can do 10 feet farther than your present office network drop:
10 - Quake/other FPS (uh, only on your lunch break ;]) :])
9 - AIM/ICQ/IRC/ABC/XYZ
8 - Internet.."research". (email)
7 - Internet.."research". (/.)
6 - Internet.."research". (pr0n...uh, only on your lunch break
5 - Internet.."research". (build up ur homepage)
4 - Editing office reports (your resume)
3 - Editing office reports (your letter of resignation)
2 - Doing productive work. (Churning code out)
1 - Pretending to do productive work. (Staring at the beautiful code you've churned out)
That's right, way more than half my "smart kids" class had to leave Western Australia to find challenging and/or financially rewarding work.
They don't mention any sort of range. They also dont say if it requires line of sight, although since its an 802.11 family standard, I would have to assume that it does not. But could it _benefit_ from line of sight ?
Given that you can build LOS point-to-point microwave xcvrs and get a few kilometers out of them, how well would this lend itself to effectively making wireless DS-3s ? I know it'd be much cheaper to put up two towers pointing at each other than it'd cost to pay for a full-clock DS-3 for even 1 _month_
Nothing is said about antenna's or anything, so while its nice that they've got lots of functionality on these small chips (233 and 68 pins, iirc), if i need a big antenna it wont be a very nice device afterall.
Finally, they mentioned that the R-5M or whatever chip supported half-duplex operation. Is this the norm in wireless ? A 54mbps half-duplex pipe seems like it could be a lot better on paper than in practice.
My opinions are my own, and do not necessarily represent those of my employer.
Manor is a big house, the one the squire (or ladies and gentlemen) live in!
Not to be rude, but is there any reason there's been so many Aussie stories on Slashdot as of late.
Not to be rude, but is there any reason there's been so many US-only stories on Slashdot as of the past three years?
-- Post No Gravy
The article says it was sold for 500 Million dollars (Australian) which converts to aproximently 259,550,000 American dollars. So, the Slashdot article should have said for "quarter of a billion dollars". Still a lot of money.
Sig goes here
The real obstacle for the take-up of wireless LANs is something that's been around for quite a while and probably isn't going to go away any time soon: government legislation.
Here in the EU (I'm in Britain but most such legislation is controlled by the EU now) there is very tight legislation on what can be broadcast without a license, even over a very small distance. With all the concerns over mobile telephone radiation, power-line radiation and whatever else I really can't see this changing anytime soon.
-- Piracy is a vicitmless crime, like punching someone in the dark.
Blaming GW Bush for the Iraq war is like blaming Ronald McDonald for the poor quality of food.