For many years now the debate on telecoms in Australia, especially when it comes to the National Broadband Network, has been framed around what we need now. This attitude has lead to some shockingly poor decisions being made such as the “rural access” programme that gave rural users 2x 64Kbps ISDN lines & called it broadband, or, if you couldn’t get ISDN, a 128k one-way satellite connection with a dial-up upstream.
Those were the days, when investment was off the cards, & Ziggy was CEO of Telstra.
The problem with this kind of thinking is that it leaves future generations with a deficit, of the bandwidth variety. We’re not alone in this problem, much of Europe, the US, & Asia suffer from the same problems, & looking at Ookla’s Promise Index, countries at 25Mbps or below average start at 23 on the list. Australia is 52nd, just behind Kazakhstan.
Before I continue, it’s interesting to note that there’s little correlation between speed & cost per Mbps, if anything it’s cheaper per Mbps to run faster services. Although much of the price differences can be put down to pay level differences in various countries.
The problem with a bandwidth deficit is that there’s no other option but to throw money at the problem, working in ICT for so long I’ve come to the conclusion you have two ways of doing things: gradual consistent upgrades, or waiting until breaking point & throwing every cent you have at the problem. I’ve seen companies opt for the latter time & time again, yet I’m always asked “why is it costing so much?”. My answer is usually rather blunt: you didn’t listen to people 3 years ago & opted to penny pinch.
Needless to say that usually garners a furrowed brow & a lot of muttered words.
So what happens when you need to throw money at the problem? Well, generally by this stage a good manager/CTO/etc will cut their losses & say “do it right this time, & make sure we can upgrade it easily later”. You don’t hear managers who are deep over their head saying “fuck it, just patch it up & hope that we can find a solution in 3 years”. Well, you do, but in 6 months the system collapses & heads roll.
I digress, the key point here is that with ICT, resurrecting old equipment or code is far from smart. Architectures move on, the needs of today are never the needs of tomorrow, so patching things up is always a very short-term solution. I’ve seen these short-term solutions extended past their usable limits many times before, I’ve seen money thrown at problems in silly ways, but it all ends up at the same point: a proper upgrade is required eventually.
Looking at telecoms, we can see the worst kind of patch it up attitudes at play. This is not just limited to Australia, but the UK & US have similar problems. Even with the BDUK (Broadband Delivery UK) project in full swing, it’s clear that the patch up isn’t sufficient to deal with current, let alone, bandwidth deficits.
The problem with only building for today is that while your network stagnates, fidelity of data goes up. This has been shown repeatedly to be the case in the past. Recent news from Netflix of their new “Super HD” (essentially 1080p HD with less artefacting) & 3D streams being opened up shows exactly how rapidly data use goes up. Super HD weighs in at 7Mbps, & 3D at 12Mbps. A traditional Netflix HD stream is only around 5Mbps. Netflix’s basic 4K TV streaming is set to start at 15Mbps, this is equivalent to the “DVD” streaming for true 4K. In reality, true 4K streams are anything from 60Mbps to 3.5Gbps, depending on how much compression is used.
THE COST OF KEEPING COPPER
ISPs the world over are already changing their attitude toward copper, either via self realisation, or legislative/public pressure. What are they ditching copper for though?
Well, I can’t say I’ve seen anyone ditch copper for a wireless style network. Sure, Ericsson & Samsung have shown some mind-blowing speeds on their MIMO channel stacked “5G” prototypes. Very nice, in a sealed environment with one user. I bet I can get my 1Gbps LAN at home to that speed, wait, I already did. It’s really just a nice tech demo, not exactly the here & now.
By the time they even get close to releasing this stuff home fibre could be pushing 10Gbps for the top users.
It seems that many are rallying against wireless networks replacing copper, with, until recently, residents of Fire Island post Hurricane Sandy being left languishing on Verizon’s “Voice Link” service. It’s essentially a mobile phone base station hooked up to a normal phone so the calls go over the mobile, rather than copper network. We used to use this at Telstra when I worked there, they were called an “interim service”, & were only given to people who had massive cable faults & only for a short period. Well, they were supposed to be, but often people could be on them for 6-12 months in new housing estates due to developers not laying copper until the estate is full.
It wasn’t until legislators became involved that Verizon quickly canned the idea of moving everyone to Voice Link services, opting to deploy FTTP to the whole of Fire Island.
In places like Africa we’re seeing FTTP deployed almost everywhere, from South Africa to Kenya, Rwanda, & even the Princes of Nigeria get a look in, not for bandwidth now, but for bandwidth in the future, &, the most important part, cost.
It’s estimated that in Australia alone the cost savings of deploying a pure fibre network is in the $600m to $700m ballpark. Nothing to snort at. That’s almost half the estimated $2b required to currently run the copper network. The problem with FTTN is that the best part of the network (main pairs, rarely worked on) are replaced & the pairs with the highest failure rates (distribution pairs) are kept.
This is without mentioning the massive increase in copper theft of recent years. Even train signalling, with its higher voltage lines, is being stolen. It’s becoming so bad that Metro in Melbourne started having helicopters with infra-red cameras patrol the lines to catch copper thieves in the act.
THE NEW BASELINE
It’s becoming more apparent that 1Gbps FTTP is becoming the new baseline, with providers across the US, Europe, Asia, & South America, opting for 1Gbps capable FTTP. The only people holding back on speeds are leviathans like BT, Verizon, & AT&T. The latter starting to wake up after Google gave them a huge kick in the arse.
The key is, the baseline speed of the services is being pushed to 1Gbps, something that’s never been discussed at great length here in Australia. Sure, we’ll have 1Gbps services for some, but will they be as cheap as the 1Gbps services in France (€38/mo for 1Gbps)? I doubt it, we’re not exactly the cheapest country to get Internet access in.
Looking at FTTP around the world, it’s picking up steam as the arguments from 5-10 years ago that “fibre is too expensive” just doesn’t wash. If a country that’s constantly at risk of insurgency, such as Pakistan, can offer 100Mbps FTTP, why are we opting for a maybe 100Mbps FTTN network?
Not everything is doom & gloom, these ads for Japanese ISP So-net’s (Sony’s ISP) 2Gbps NURO FTTP service are quite awesome. Enjoy: