It isn’t merely NBN-detractors who claim that wireless internet is the way of the future. Everyone loves wireless. But does it have a future?
You’ve doubtless seen the figures. Wireless usage is increasing nearly exponentially and by 2020 could reach 1Gb per day per user. One report estimates that roaming wireless usage will generate $80bn in annual revenue by 2017. Sales of smartphones, tablets and phablets - there’s still a slight chance that name won’t catch on – are roaring along unchecked. We’re constantly told that the ability to work while on the move is just the thing, although I’d like to see some figures on how much productivity is lost through mobile workers suffering injuries, if only from colliding with each other.
Anyway, I’ve spent some time working as a writer at the Australian Open throughout the tournament this last week and a half, and mobile working has been great. I can’t imagine how journalists covered such events in the olden days (like ten years ago). Actually I don’t have to imagine it, since the constant failure of wireless data on the AO grounds recreated it perfectly. It was less fun than I'd hoped.
Anyone who has visited the Australian Open tennis in the last few weeks will have obtained a firsthand demonstration of the fundamental shortcoming of wireless data. When lots of people are trying to use the same network, performance is affected, because wireless is a shared medium. At peak times there were 40,000 people on the grounds at Melbourne Park, and although only a small proportion of those would have been trying to access Telstra’s mobile networks at any one time, the number was still high enough that data performance was either hopelessly slow or non-existent. Twitter was unusable, which was a real problem for those keen to tell hordes of near-strangers precisely what they were thinking or eating at any given moment. iMessaging on iPhones was spotty at best. The Australian Open’s own app, which provides allegedly real-time updates to scores and schedules, was basically useless.
However, those lucky enough to have a media accreditation to the event are able to use the tournament’s dedicated Wi-Fi service, which links directly in to the fixed network. Now, I won’t claim this service was perfect. Wi-Fi went down for an hour or so yesterday, leaving a very large number of foreign journalists in a sorry state, since their option was to remain silent for a while (unthinkable) or run up roaming data costs (sacre bleu). Last year the Wi-Fi went down for an entire day. But in both cases the issue wasn’t one of congestion, but a technical problem on-site. Unsurprisingly, the fixed network had ample capacity – it is the premier sporting event in Australia, and Melbourne Park has some pretty fat pipes running into it.
I’m not going to into the shortcomings of wireless solutions here, and why fixed data connections will remain fundamental for many years to come. I want to look at the idea of whether the wireless networks can even support the amount of customers determined to use them in the near future. After all, if your smartphone or phablet (honestly, is there a petition I can sign to have that name banned?) can’t access the internet, it becomes considerably less useful. The experience of users at Melbourne Park is one commonly experienced by users in capital cities around Australia and the world. Congestion is having a serious impact on network performance. Early LTE adopters on Optus and Telstra have noticed that performance after only a year has shown noticeable degradation as more subscribers have joined.
According to a new report by Deloitte, this isn't a temporary issue that will be fixed by building more capacity, but a fundamental issue caused by everyone trying to wirelessly connect at the same time, and it’s only going to get worse. The average smartphone drives 35 times more data than a regular mobile phone that simply makes calls and sends SMS, and as more and more people switch to the more powerful devices there will be an increasing ‘spectrum shortfall’.
They predict that global roll-outs of LTE will only delay the issue by about a year, and argue that because wireless data is now an increasingly scarce resource, it should be priced accordingly. By that analysis, the recent reserve price set by the Australian Government for its spectrum auction makes a lot more sense. (On the other hand I'm probably being generous to assume Senator Conroy based the price on anything more complicated than a directive to generate revenue.)
Anyway, what does this mean for the increasingly mobile work force, who (we’re told) needs all that data in order to conduct business properly? What can you do to change it? The answer is nothing. There’s nothing you can do. It’s just the way it is. Wireless spectrum is a limited resource, and the idea that an unlimited number of people could access it was always flawed.
There are of course ways to game that system, such as by running more fixed capacity in to service more base stations. Not unlike the media Wi-Fi at Melbourne Park.
UPDATE: CNet has released figures revealing that 4G data speeds have dropped by about 28% since October 2012, due most likely to an increasing number of users moving on to the network. 3G speeds have remained fairly stable, which is to say uniformly low.