Point-to-Point Mobiles

or: How to cost a telecommunications company 10 Billion dollars.

Why do Mobile phones always have to go through the exchange, even when they are a bare few metres apart?

In my household, there are two people with Mobile phones. This leads (occasionally) to dirty tricks. One person dials the other's mobile while sitting in the same room. Then, as the person is lifting it to their ear, you hang up. They shout 'hello' into it for a while, and shake their head. Then you do it again. And again. Serenely sitting there with your hands in your lap, watching them, trying to keep a straight face. But, think about what's happening: One mobile is powering up, transmitting for every watt it's battery is worth to a local cell antenna, which then routes back to the other mobile, activating it, and again causing massive battery drain. If the connection actually goes through, you then get charged exhorbitant rates for the privilege. Why on earth should it work like that? The two phones are not seperated by more than a few metres. Let's change the system to be like this: Before going all the way to to local antenna, the calling mobile transmits a short 'Are you there?' burst containing the destination mobile number. The recieving mobile picks the signal up, and returns the burst. Both mobiles have now established that they are within transmission distance of each other: They don't need to go through the exchange. One can transmit straight to the other. Point-to-point mobile communication. Just like a walkie-talkie. The average Digital Mobile phone is an astonishingly sophisticated piece of computer and communications engeneering, capable of performing all sorts of tricks. Of course, "smarter" is synonymous with "cheaper", but which Telecommunications company really wants its customers paying less? What is needed for this to happen? First, the software in the mobile phone needs to be capable of performing without the aid of a mobile base station. This isn't hard - mobile phones tend to contain fairly sophisticated little microcontrollers, capable of encryption, authentication, handshaking... a little routing shouldn't tax it too much. A more serious issue is transmission frequency. Digital Mobile Phones in Australia are based on the GSM standard, which defines two frequency bands: 890-915Mhz for transmission from the mobile to the base station, and 935-960Mhz for transmission from the base station back to the mobile. Each 'channel' is 200Khz wide, and the transmit and recieve channels are locked 45Mhz apart. Each mobile uses 1/8th of the channel in rotation with 7 other phones. This means two things. A standard mobile phone cannot transmit and recieve simultaneously on the same frequency, and unless the mobile phone has a software adjustable frequency range greater than 45Mhz, then there is no overlap between transmit and receive bands. Something else that is interesting is the extreme lack of bandwidth available in the mobile system. There are only 125 duplex channels available in the band, and with 8x time division multiplexing, only 1000 mobiles can be in use at any one time. Of course, the Cellular system gets around this by re-using frequencies amongst cells, but doing this, you get limited to 1/6th to 1/10th of the total channels in each cell. So, only around 120 simultaneous connections are available in each cell. Salvation arrives in the form of the 20Mhz gap between the bands. 20Mhz is a lot of room: enough for 800 channels using the existing, relativley inefficient system. And given that we can probably re-use those frequencies after each kilometre, we get a lot of available channels across the average city. 4 million simultaneous connections, according to my numbers. As far as the 45Mhz xmit/recv gap goes, we don't need true duplex action. If both mobiles are going to be communicating on the same frequency, then they have to be constantly switching between xmit and recv anyway. So, what do we need in a mobile phone? We need one that can transmit over a band of 870-935Mhz, making it's 45Mhz coupled receive band span 915-980Mhz. This gives a useable overlap in the 915-935Mhz band, sandwiching nicely in the gap. By extending a Digital Mobile's transmit and recieve frequency ranges by 20Mhz each way, an overlap is created in currently unused spectrum where any two mobiles can communicate. If that's the case, then all that is is needed is a software change to your mobile. Yup. Change the EPROM. $5 worth of hardware, and probably $50 for the technician at murderous rates. Now, think about what that means. First, it changes some of the characteristics of how you can use your phone. The closer the person is, the less power it takes to talk to them. This means more talk time for your battery. Next, if someone's in the same suburb, then it costs nothing to make the call. Take these two into combination, and you end up using it instead of an intercom when in the same house. And, in outback regions that have no mobile coverage, you can still contact people nearby. Mobile usage skyrockets. As you get further away, you get to make the decision about whether you continue with the point-to-point for as long as possible (draining the battery faster the farther you travel) or take the hit to the wallet at some prescribed distance when the costs of going through the exchange override the inconvinience of a flat battery. Past a certain distance, you of course have to go back to using the exchange. (Or maybe not. I'll develop this later.) You phone can handle all this automatically, because it's got a built-in signal strength indicator. So, with a software hack in your phone, you get free in-suburb calls. Let's add another fact. Most Mobile phones are now Digital. What else is digital? Lots of things. Lots and lots of things. What if all those digital things could communicate with each other over short distances for free? Well, bye-bye office LAN wiring hassles for one. I can file transfer to my friend a few doors down. EFTPOS and Teller machines can communicate directly to their local bank branches. As I pass shops in the evening, I can read mobile numbers off the windows, call to get their catalogs, and let them know I was passing but they were inconviniently closed. While sitting at a bus stop, I'll call the number to hear the timetable. As I wander around the huge campus of my University, I can call my friends and ask them where they are. Now, take another fact into consideration. Using the signal strength indicatior in my phone and a little clever logic, we have a rangefinder. Once the privacy issues are cleared up (my phone only allows other phones to locate it when I've explicitly set the allowable numbers, for instance) then my phone will beep when my friends are near. (Great if I'm desperately trying to find their house, or we've all agreed to meet near some landmark but we're doing the "maybe he's over there" shuffle.) My office can tell when I'm in. (and route intra-office calls through to me) My house can tell when I'm at the door, my car can accept alarm codes, I can track my way to the nearest pub, I can press an emergency button and the police can find me, I can listen to music, I can page a passing Taxi. All these things will enhance my life and piss off my phone company. But I'm quite sure that if I bought two mobile phones and started hacking the EPROMS, then Telstra would have the FCC down on my ass faster than a techno beat track We need to stop thinking in terms of the telephone/exchange model, and more in terms of a radio/repeater. Don't say "Mobile phone", say "digital walkie-talkie". You'll find this changes you outlook. I'm thinking about some concrete examples that might be illuminating: Some office environments would be revolutionised by cheap point-to-point network services. Workers with Laptops and a mobile phone become truly mobile within their offices, and can set up wherever they may be needed. All calls can be routed from the switchboard to their mobile, and their laptop can exist on a wireless LAN. Also, a little work with the rangefinder would mean that my co-workers can find me, or the automated office system can tell if I'm in, or out. I know some Archaeologists. They tend to work in teams in really out-of-the-way places. Point-to point mobiles would improve their work, and enhance their safety. The same applies to Bushwalking parties: Not only can everyone stay in contact, but if someone gets lost, they can be quickly located. If you're traveling long distances by car, and you break down in the middle of the night on some god-forsaken country track outside mobile coverage, then your phone, oh-so-useful in similar situations in the city might just be good enough to get you out of trouble if you could broadcast a 'help' signal to anyone else in range. In a natural disaster like a cyclone, which they all-to-often get in tropical Queensland, or an Earthquake, which has happened in Darwin, or someone sabotaging the network with an axe, like in Sydney, then mobile phones will still work! People buy mobile phones for safety, just as much as for convenience. For the network to experience outages in civil emergencies, like the cell tower being taken down by an earthquake, is not only inconvenient but life-threatening. Point-to-point mobiles will save lives. Now, what was I saying before about not even needing the exchange for middle-distance calls? Have you ever heard of packet radio? I'll explain. Packet Radio was created by some Amateur Radio (Ham Radio) boffins who wanted their computers to talk. They essentially wired up modems to their radio rigs. Now, one of the drawbacks of shortwave (or CB, or just about any kind of amateur radio) is that it's simplex. You can either be transmitting or recieving, but not both. Hence the little push-to-talk doodads you get on the microphones. What this meant is that they couldn't have their modems blurting funny noises into their radios all the time, as then they couldn't hear anything coming the other way. So, they hacked with TCP/IP a bit, and used the idea that you can break up a longish message into smaller packets, and send each one individually. In this case, the computer/modem/radio rig would listen to make sure the air was free, and then quickly blart out one packet over the air. Then, it would stop, listen, and make sure that there hadn't been a collision and that it got a response back from the far station. Then it woulds send the next packet. So, a packet radio station, when active, would go 'blart, listen, blart, listen, blart' ad nauseum. Packet radio breaks data into 'packets', which are transmitted by radio. Digital Mobile phones work exactly the same. Now, here's the really tricky part. You, with your radio rig, might be sitting midway between two people who can both communicate with you, but not with each other. And, being a helpful citizen, you might want to relay the messages. This is called, easily enough, repeating. In the CB days, people would relay the messages by voice. With packet radio, you could leave the computer rig set up, and it would do it automatically. A network of packet enthusiasts developed. Long-distance transmitters were set up and a trans-pacific link to the US was created. A parallel internet developed, running the same protocols. Soon they all became the same thing. One man sailed around the world, making infrequent landfall, but constantly updating a usenet newsgroup of his progress, all over packet radio. It's all still there! This public spirit of like-minded individuality made them pool their resources together to create a national radio network of packet-flinging computers. At no cost to themselves apart from a little bit of electricity. Packet relay software was developed that re-transmitted other peoples data while your machine was otherwise 'idle'. When everyone co-operated, a worldwide network was created. Now, how many mobile phones are there? At any one time, what percentage are in use? How many are sitting in charger sockets, doing "nothing"? Remember, all mobile phones are little computers with a digital radio transmitter/reciever built in. Then think to yourself: If you call someone on the other side of the city, how many inactive mobile phones are there between you? If the maximum point-to-point distance between them was 500 metres, how many hops would be necessary before you got to where you needed to go? It's not long before you start thinking of mobile phones in terms of little packet radio repeaters, with a bandwidth wide enough for real-time voice. If the population of mobile phones becomes dense enough, then there are always adequate numbers of repeaters available, all in short distance. Your mobile is now always active, true, but now it's always on at 1/10th the power, rather than infrequently on at full power. And you have no phone bills. Worried about equality of use? Build it into the software. For every hour you spend on the shared network, your little packet repeater gives the same back to the community. It can probably even time-share while you're talking: while you use it, everyone else can use you. A bit of statistical analysis shows that once the number of people with these phones grows past a certain size, then the network gets at least as reliable as the network already in place. You thought that point-to-point mobiles would scare a telecommunications company! How about free mobile calls across entire cities? Depending on topology, we'd be left with the long distance carriers, and that's about it. So, let's review the compelling reasons why mobiles should work point to point, even without packet forwarding: Reduces phone battery drain when talking to people close by. Allows economical use of your phone to close-by destinations, including digital services like Wireless LAN's. Gives "Community Use" applications. Provides redundancy in the case of civil strife - union actions, network outages, natural disasters or sabotage, and can save lives. Enables use in country regions that do not have mobile coverage - and gives a safety net for long-distance channels. Provides a greater level of emergency assistance in life-threatening situations. Now, let's look at the objections that people will raise against the scheme. It's too costly Rubbish. All that needs to change is the software, and to extend the on-board software controlled frequency generator by 40Mhz. It would interfere with the existing network. No, because the point-to-point phones use unallocated spectrum between the existing bands. It would interfere with other spectrum uses. This is somewhat true. But we have to take a path of greatest public good. An uncontrolled system would degenerate into chaos. Tell that to the internet. Bzzzt. Wrong. You can't garuntee service. Even if service can't be garanteed, at least we're not paying for it. There's not enough spectrum to hold all the calls, especially if people establish 'permanent' connections using cellphones. Yes, there is. Even using the horribly inefficient round-robin scheme already in place, there's enough space for 400 conversations per square kilometre with all mobiles at a maximum permitted strength. Using a more intelligent scheme, there's probably enough for 1000 conversations per square kilometre. That's enough for 5 million conversation in a city the size of Brisbane, with the lowest availability at 80,000 conversations in the CBD. The govenment won't allow it. Er. Ok, you got me on that one. And that's the reason it will fail. I invite you to look at the Spectrum Management Agency homepage. In particular look at their PTS Apparatus Licence fees and see how much the Telecom companies are paying to use bandwidth. Getting more spectrum for this kind of public use inside the existing cellular bands is going to be a beauracratic nightmare. I was interested to find out that all Australians pay an Spectrum Tax. Yes, we're taxed. For spectrum space. Regardless of whether we use it or not. (But to be fair, we all use it to some degree) However, I think this gives us the right to ask the government to reallocate or assign spectrum space for this kind of public good. With this re-organization, all it takes is a working panel of wireless network specialists to develop a communications standard (probably modelled on Packet radio and TCP/IP) and one single mobile phone manufacturer. (who would suddenly not be able to make his phones fast enough) Orinoco

Coverage There are three major digital mobile phone companies in Australia: Telstra, Optus, and Vodaphone (who does not have a home page outside the UK, and not a very good one in it). Telstra have promised to send me a coverage map, by smail mail, (they can't fax it, you see) and apparently it will be here in two weeks. Optus was slightly better: Theirs should be here in four days. Look back here for when it arrives. The SMA The Spectrum Management Agency is an organization I hadn't even heard of until today, when I found out that they are taxing me. This is called the Spectrum Access Tax and, as far as I can determine from their Financial Statement, they cost us $36 Million each year, which is a sizeable portion of their income. The other chunk of revenue is composed of Licencing fees, MDA application fees, and the GMS surchage which adds up to about $234 million. This cost, of course, eventually gets filtered back to us anyway in the form of advertising (which becomes a product cost on everything we buy), mobile charges, and other taxes. This means two things: (1) That selling spectrum is a very lucrative proposition. (2) That the spectrum is ours, because not only does the SMA specifically define it as a 'a community resource', but we annually pay for it as well. We own, and rent it! I personally consider $2 per year for every man, woman, and child in Australia to be well worth the cost of a properly managed spectrum. Of course, if you count the transmitted cost from selling the other services, it goes up to something like $16. GSM Securitry A couple of people are asking questions about the security of the Digital Phone encryption software. It seems that it's well within the bounds of an average workstation to crack the A5 cypher used in a few weeks, or for dedicated circuitry to do it in half a day. See this article for the details More disturbing is the report that, I quote: "there seems to be an even faster attack. As the clock control is stop-go rather than 1-2, one would expect some kind of correlation attack to be possible, and on June 3rd, Dr Simon Shepherd of Bradford University was due to present an attack on A5 to an IEE colloquium in London. However, his talk was spiked at the last minute by GCHQ...I hope that placing A5 in the public domain will lead to the embargo on Simon's paper being lifted." Information on the Xylinx chip mentioned can be found here. Speech Encoding One of the more interesting technologies present in Digital Mobiles phones is the lossy speech encoding algorithm which gives a constant 9.7:1 compression ratio. The site includes UNIX source to which implements the algorithms, providing easy tools to compress voice. There are even versions for the PC and mac. With a little help, GSM 06.10 could easily become a de-facto standard for voice compression. L0pht Much of the information for this page was culled through the techo-savvy links of L0pht Heavy Industries. A highly recommened site. Wanted This is the section where I ask for reader help. Hopefully someone out there can answer my questions too. I'm looking for real digital mobile coverage maps of Australia. Telstra's advertising says it has 90% mobile coverage, but if they're talking about population coverage, which I think they are, then that's only about 4% of the country by area. I'm sure that there are some mobile phones out there somewhere that already have the ability (although unused) to transmit and received in those between bands I talked about, and which could be made to work Point-to-point just by hacking the ROM. If there are any digital phone hackers out there who know for sure, I'd like to find out which models by which manufacturers are capable. And if anyone were to, uh, 'modify' phones in this manner, which I of course do not condone, then I'd be interested in hearing about the results.