Mobile demand might outstrip capacity within 12 months
- Optus will seek to re-farm old spectrum for faster data
- 700Mhz spectrum to go on sale this year, but won't be usable until 2016
- 4 times as much radio capacity available now than in 2008; but Aussies use 30 times as much data
The Wall Street Journal carried a surprisingly frank op-ed from Randall Stephenson this week, in which the AT&T CEO publicly fretted over dwindling mobile broadband spectrum in the face of more and more mobile services.
Stephenson was talking about the US market, but the situation in Australia is possibly even more dire. Right now, there is about 4 times as much radio spectrum available as there was in 2007, before the iPhone debuted in Australia, and before mobile broadband became a widespread phenomenon. But we use about 30 times as much mobile data as we did back then – and that’s increasing rapidly.
Not surprisingly, the gist of Stephenson’s article is a call for laxer regulation in deploying wireless infrastructure, but he also raises the issue of spectrum speculation. This is where companies with little to no actual telecoms business buy up slots on the radio dial, so to speak. They then hang on to those licenses and sell them to the highest bidder when the situation gets desperate.
Speculation in Australia will be limited when the 700Mhz band, due to be auctioned off this November will be restricted to known network operators – Telstra, Optus and Vodafone have already expressed their intention to compete, with Google a possible dark horse contender (Google has been getting into the telecoms game, deploying a Fibre-T0-The-Premises network in Kansas City). The 700Mhz band, previously used for analogue TV transmission, is considered something of a prime candidate for data transfer- it provides the best compromise between throughput per-second and propagation (strength of signal over distance) currently available to RF engineers. Anything lower dips into frequencies used for emergency services. But the 700Mhz band won't be operational until 2015 at the earliest; it will still reach capacity eventually, and we're short of spectrum right now.
(ed note June 19 - The ACCC has now delayed this auction to mid 2013).
Telstra has been taking the lead in 're-farming' - using old spectrum no longer deployed for slower standards, and updating their equipment to transmit at faster standards. James Howe, from Telstra's media team, said "Telstra has been progressively refarming our 1800 MHz 2G spectrum for use in (our) 4G network". As for the 700Mhz spectrum auction, Howe said "Should Telstra secure allocation of spectrum through this auction, Telstra would use the additional spectrum to further develop its mobile network built today on world leading 3G HSPA+ and 4G LTE technology".
How do we ‘run out’?
Every form of data transmission gets back to the fundamental issue of 1 and 0. You must have a means of ‘modulating’ the data, or turning it into positive (1) and negative (0) pulses. Electrical copper cabling (ie. phone lines) achieves this by turning the circuit ON and OFF several million times a second. Fibre achieves this by turning a laser diode ON and OFF millions of times a second; Radio Frequency (RF) achieves this by switching channels. This requires more power, and of course, the use of more channels.
Lasers can also switch colours without interfering with each other – your data can come on red light, and your TV transmission could come on blue, with no sharing of bandwidth. This part gets right into the nitty gritty of physics, but as magical as it seems to the uninitiated, it’s actually quite rudimentary. Light has ‘frequencies’ in magnitudes of hundreds of millions (think of each shade of colour in the visible light spectrum, then think of UV and all the other frequencies of light we can’t see). RF, meanwhile, has about 1000 channels per band, and about 300,000 bands. But this is electromagnetic radiation we’re talking about – anything above 6000Mhz interferes with oxygen molecules, making them worthless over distances of a few millimeters.
In a summary from Karl Zingre, a senior RF engineer at the Australian Synchrotron in Clayton, Victoria; “The increase in data flow may be compensated with higher compression, improved modulation methods and other tricks such as dynamic bandwidth control. But principally, (RF data transmission) requires new bands and frequencies to avoid interferences with other users, loss of data or waiting queues. However the frequency cannot just go higher and higher due to available technologies, increased losses and costs. There is also still an unknown human risk of radio frequencies and whether or not or how much they are a health issue”.
The issue of dwindling spectrum is becoming a hot-button political issue in Australia, thanks to the debate around a National Broadband Network (NBN). In short, the government’s all-fibre network would take 10 years and be costly to build, but would make the same speeds available to 93% of the population – and that’s speeds at home, not speed from a central exchange. It would also replace the copper ‘last mile’ network that gives Telstra a monopoly, and effectively dissolve Telstra’s Wholesale arm and make Telstra an equal retailer.
The coalition’s alternative, which they maintain will be cheaper and quicker to build, would be to offer fibre to the node, or fibre to within a few hundred metres of everyone’s home – meaning that Telstra Wholesale (or an equivalent) would still have a monopoly on the last few hundred metres of copper bringing every home a connection. Speeds would not be as fast, but they’d be to a level that would still significantly boost the current average. But still, this is considered a patchwork fix that would require upgrades to more fibre over time anyway.
The fallback alternative argument against a more comprehensive approach is that fixed broadband will be obsolete anyway, by the time it’s built. With more and more Australians taking up mobile broadband using RF (radio frequency), the idea that people will want a fixed connection in 10 years time is supposedly ludicrous.
This ignores the discussion of where your RF transmission comes from. Essentially, THERE ARE NO WIRELESS CONNECTIONS TO THE INTERNET. Everything “Mobile”is just a longer range connection to the wired network. Even a satellite signal would eventually be caught by a terrestrial dish, wired to…you guessed it, a fibre connection.
Right now, telecoms companies are wiring their transmission towers with fibre at the bottom, to offload local mobile traffic through a higher capacity ‘pipe’. But even getting that fibre capacity into the hands of users would require high-capacity radio signals. And the higher the capacity, the shorter the range.
Iron and Dinosaurs
It’s kind of like the conundrum that we have with Iron. Iron is strong and hugely plentiful – but it turns useless in oxygen, which is where we mostly need it! So we alloy it with other elements to make it rust resistant. That compromises its strength, but makes it useful.
Similarly, we need radio to eliminate the problem of distance. But the further a radio wave reaches, the less it can carry (and penetrate walls, for that matter). Radio waves with similar capacity to fibre would carry over a few…millimeters. And even ‘boosting’ that signal would require nuclear levels of energy. Also a concern - extremely high frequency radio waves are also used in advanced weapons system to, you know, kill people.
So instead we rely on safer, lower-frequency (lower capacity) radio waves to send along data at larger distances, but that means that only little bits of data can be sent. RF is akin to dinosaurs – the more powerful it gets (Tyrannosaurus Rex!) the shorter its arms get.
How we’ll actually use RF in the future
You say tower, I say 'macrocell. Source: Reuben Bowles
Current mobile data technology is split into cells. A macrocell is the one most people associate with ‘wireless’ broadband – this would be a mobile transmission tower, providing a blanket of cover to the area. Macrocells can operate on several frequencies. The most common to now has been the 2100 MHz frequency, used by Optus, Telstra and Vodafone to provide 3G data and voice over distances of about 10 km; the 850 Mhz network used by Telstra’s NextG and the newer parts of Vodafone’s network can reach further - but carries less data per second. But what's the point of carrying lots of data if it peters out by the time it reaches your mobile phone?
Wi-Fi is a form of wireless connection with much higher ‘capacity’ – it carries much more data per second, but only over a few metres distance (up to a kilometer with newer standards, but more like 10-20 metres). Wi-Fi operates on 2 frequencies, 2400Mhz and 5000Mhz, which are of course higher frequencies than what Macrocells use. It is usually used for home based connections, and are usually fixed to a direct line (fibre or copper).
So the future will be a mix of Macrocell and Wi-Fi, with the higher capacity Wi-Fi being used to ‘off-load’ heavy traffic to make Macrocell more available for voice and small data. The move towards mobile apps, which tend to sip bandwidth and use sophisticated software to lay out data more efficiently, may help here. Checking your tram timetable app, or making a call, can be routed to a Macrocell; downloading a song or a movie on your laptop while at a café can be routed to a local Wi-Fi transmitter. But this would require much smarter software on the ISP end, and ubiquitous fibre under every street post.
On this issue, Telstra's James Howe added "Telstra no longer deploys WiFi hotspots. We are however constantly examining technology roadmaps and are considering business models that use WiFi for offload and delivering a service integrated with cellular networks".
Luckily, both the current government’s solution and the coalitions approach should account for this. Fibre-To-The-Node could incorporate a system that sees every “Node” (a utility pole every few hundred metres) carry both a Macrocell antenna and a Wi-Fi repeater, both hooked into a fibre base. Since most devices – laptops, desktops, tablets, smartphones, Smart TVs, printers, fridges, toasters – come equipped with Wi-Fi receivers, then all of these can use the neighbourhood Wi-Fi signal for high capacity data transfer. For everyone just driving through, their device can connect to the Macrocell. This would require big advancements in Macrocell technology (ie. making them smaller and cheaper) and there’s still a height issue – Macrocells are usually up high for a reason. But it could be done.
Of course, with fibre running to every streetpost, you’re getting into expenses that make it similarly priced to run fibre to every home, which is the end-point of all of this. Fibre to each home is overkill right now, but that’s kinda the point – it would last several generations before needing a superior technology. And it’s hard to even conceive what that technology would be. Silicon, the base material of fibre-optics and the MEANS of transmission, is plentiful and extremely durable. And light, the METHOD of transmission, is kinda the end-point of the universe. The only step would be to shoot a data laser, through the air, right into your home – and that provides its own barriers to entry.