|
|  |
| |
|
| | The Telecom Network for Professionals and Users Network is not currently active and cannot accept new posts | | How much can you put in a 5 pound bag? | Views: 1211 | | Apr 04, 2007 6:30 pm | | How much can you put in a 5 pound bag? | # | 
Ken Hilving | | An interesting article about Moore's Law, and more importantly the IC.
Watching wireless push the amount of data available within a particular chunk of spectrum, I wonder how long before a similar problem throttles new development in wireless services.
http://www.smh.com.au/news/technology/for-moores-law-the-end-is-nigh/2007/03/21/1174153147545.html
Of course, there may be some new approach, new technology just on the verge of becoming viable. Anyone care to venture a guess on what it will be?Private Reply to Ken Hilving | | Apr 05, 2007 2:38 am | | re: How much can you put in a 5 pound bag? | # |
Scott Hinojos | | I'm not going to pretend to know anything about radio frequency, but does anyone think something like wave division multiplexing (WDM) could be used with RF?
Another question I have is, when a carrier purchases a frequency, say 850 MHz, is it /exactly/ 850, or is 850, give or take 50-100 MHz? If it's the latter, could wireless receivers be made to accept mutilple streams? If a carrier uses an extremely coarse WDM (10-50 MHz), they could widen the throughput exponentially. Hey, it works with fiber. :)
If nothing else, moving to higher frequencies is a stop-gap solution, like with the 2100 MHz band. I've seen a few 5-band GSM handsets out there for a year or more...
Is anyone working on mobile WiMax?
Private Reply to Scott Hinojos | | Apr 05, 2007 3:05 pm | | re: re: How much can you put in a 5 pound bag? | # |
Ken Hilving | | The RF equivalent of WDM is FDM, and it has a long history.
A key difference between wireless and wired (copper or fiber) is that the signals are not constrained in wireless.
If we run out of capacity over a wired network, we can add or activate additional paths fairly easily. The new paths can use the same frequencies (waves) and modulation techniques, allowing us to double our capacity. This can be done time and again along a given physical path until we run out of physical room for new wires. We can then expand that room (new conduit) and press onward. This is a huge potential growth ceiling, especially when using fiber.
In wireless, we have to worry about interference. For a given frequency we can only use it once within a geographical footprint. That footprint is based on signal strength, and we have to be far enough beyond its range before using it again or we have interference.
Operating in free space, we also have to consider environmental factors. Certain frequencies are effectively blocked by certain materials such as plants and buildings. Just as you block visible light but x-rays pass through you, various materials block different frequencies of RF.
We also have a greater concern for interference from other RF noise. For example, solar flares can disrupt certain frequencies. http://news.yahoo.com/s/nm/20070404/sc_nm/space_gps_dc_1
In addition to these issues, we have the other uses in the radio spectrum. We have traditional radio and broadcast TV, GPS, satellite radio and TV, HDTV which demands more spectrum, two way radio communications, radar, RFID, and more.
While we have made significant improvements in our ability to transmit and receive within a narrow band, and this has increased the number of discrete signals available, we still face a ceiling within any given geographical area. That ceiling is based on signal strengths, footprint, and blockage.
So what's it all mean? There is a limit to how much can be transmitted using wireless within a given area. As we increase the amount of data per user, we decrease the number of possible users at a given time.
My first question is where is that limit and how close are we to reaching it as we move to 3G and a more balanced traffic flow from user generated content?
My next question is what solutions to expanding that limit are on the horizon?
For those of us who remember the early days of computers, watching Moore's Law approach its limit in silicone technology is incredible. Given the rate of innovation and commercialization of wireless, I wonder if I will see that maxed out as well. Private Reply to Ken Hilving | | Apr 06, 2007 12:53 pm | | re: re: re: How much can you put in a 5 pound bag? | # |
Tom Foale | | There are theoretical limits and, similarly to Moore's Law, the boundaries of what is achievable keep getting pushed back. I was at a conference last week where I was shown a graph similar to a log chart of Moores Law, but for radio. It had four parallel lines - one for wide area mobile, one for metropolitan networks, one for LANs and one for PANs (personal), with very short range having the highest rates. Each of these is still increasing at a relatively stable rate. We were also shown some research and standards development work on Gigabit WiMAX, which would hit the next spot on the MAN line if it is delivered on time.
New technologies keep adding to the capabilities of wireless, like OFDM and MIMO, and there are more areas for improvement. The efficiency of our WiMAX network is about 3.5bits/Hz and with further development WiMAX should be capable of delivering 4.5bits/Hz. This means that with a 10MHz channel we can deliver around 20Mbps of customer data from a single sector, exclusive of overheads.
The way that all radio technology works is by modulation of an RF carrier by an IF (intermediate frequency) signal. This results in two identical sidebands either side of the carrier frequency, one of which is filtered out to give a band of frequencies from which the signal can be recovered by a receiver. This makes spectrum a scarce resource because two different users of a band can interfere with each other, so spectrum tends to be allocated in non-overlapping bands to single users - or at least to non-interfering users. Although modern technologies such as WiFi can share the same spectrum quite effectively, they do so by avoiding transmitting at the same time.
There is one company with an intriguing new idea. What if, instead of one RF carrier and an IF modulator, we used a large number of harmonic RF carriers and switch them precisely at a common zero crossing point? In theory you should see a series of precise intermittent single frequencies which to most receivers would register as a small amount of noise. This couldn't have been done before now because the RF section was always the most expensive part of the radio, but modern chipsets have reduced that. It could allow many operators to co-exist in a small amount of spectrum, reducing the scarcity of spectrum. It can also transmit data at the same rate as the carrier or higher (depending only on the number of RF frequencies used) - meaning much higher data bandwidths at lower frequencies.
The company is XG and the technology is XMAX - there are patents available which can be found under the name bobier. I think the company is commercially unsound, but the technology looks interesting - if it works. It still lacks the surrounding standards necessary to create something like WiMAX or 3G. It took Qualcomm 10 years to get its CDMA technology accepted by the mobile world, so XG have taken a different route, but I don't think they will have much success with it for a number of reasons. It takes more than a good idea to make a good business.Private Reply to Tom Foale | | Apr 09, 2007 2:35 am | | re: re: re: re: How much can you put in a 5 pound bag? | # |
Ken Hilving | | Tom, I suspected the issue of spectrum capacity has been discussed in depth by various groups.
What would be really useful is a graphical representation of the current status and barriers. I am not as convinced of the long term growth. History has shown us a trend towards waste of capability - most of today's code lacks the elegance and efficiency of earlier code where memory, non-volatile storage, and communications paths were significant constraints.
For all communications, here is another solution to ponder. We have increased the internal bus from 8 to 64 parallel lines. This greatly reduced the number of cycles required to accomplish a given task. Yet we are still essentially using a single serial path for all communications. This made sense when each path was its own cable run with its own set of DCE. Is it time to move to parallel paths for communications? Private Reply to Ken Hilving | | Apr 17, 2007 5:33 pm | | Regarding xG, absolutely false | # |
Marc Dannenberg | | The commercial abiliry of xG Technology to fulfill its rollout is 2nd to none. You obviously have done some Due Dilligence by doing a simple search on Google. The company has signed on with Telefonica and National Grid, two behemoths in the communications industry. Do yourself a favor, Tom, and dig a little deeper before you post nonsense regarding the capabilities of xG´s commercial reach. The company trades at over 2 Billion dollars without making any money yet. I believe that´s a great job on the commercial end. Private Reply to Marc Dannenberg | | Apr 20, 2007 1:12 pm | | re: Regarding xG, absolutely false | # |
Tom Foale | | Marc
Actually, the commercial ability of xG Technology to fulfill its rollout is completely unproven. As we've had a private discussion on this you know my views. However, many companies that have traded at huge valuations without significant revenues have subsequently been shown to be wearing the emperor's clothes. There's a lot of execution to be done between a significant valuation based on promise and significant revenues that requires top-notch strategy and management - regardless of how good the technology is.
High valuation without revenues usually represents the triumph of hope (or hype) over experience. NGW is now private so their shareholder may be happy with this strategy, but our shareholders would be very unhappy if we followed a strategy that failed because of someone else's undeliverable technology or poor execution. XG's decision to shroud its technology in secrecy means that we could not inform our future investors of what we were basing that strategy on - so why should they invest in us? That XG decision, plus the dodgy NDA clause I showed you, does not look like a company executing effectively to me and my fellow directors.
TomPrivate Reply to Tom Foale | | |
| |
| |
