I recently read an article in LightReading titled “All That’s Gigabit Doesn’t Glitter.” The article asks the question if the industry really needs to make the leap to gigabit speeds. It talks about the industry having other options that can satisfy broadband demand but that telco executives get hooked into the gigabit advertising and want to make the gigabit claim. A few of the points made by the article are thought-provoking and I thought today I’d dig deeper into a few of those ideas.
The big question of course is if telco providers need to be offering gigabit speeds, and it’s a great question. I live in a cord cutter family and I figure that my download needs vary between 25 Mbps and 50 Mbps at any given time (look forward to a blog soon that demonstrates this requirement). I can picture homes with more than our three family members needing more since the amount of download speed needed is largely a factor of the number of simultaneous downloads. And certainly there are people who work at home in data intensive jobs that need far more than this.
There is no doubt that a gigabit is a lot more broadband than I need. If we look at my maximum usage need of 50 Mbps then a gigabit is 20 times more bandwidth capacity than I am likely to need. But I want to harken back to our broadband history to talk about the last time we saw a 20-fold increase in available bandwidth.
A lot of my readers are old enough to remember the agony of working on dial-up Internet. It could take as much as a minute at 56 kbps just to view a picture on the Internet. And we all remember the misery that came when you would start a software update at bedtime and pray that the signal didn’t get interrupted during the multi-hour download process.
But then along came 1 Mbps DSL. This felt like nirvana and it was 20 time faster than dial-up. We were all so excited to get a T1 to our homes. And as millions quickly upgraded to the new technology the services on the web upped their game. Applications became more bandwidth intensive, program downloads grew larger, web sites were suddenly filled with pictures that you didn’t have to wait to see.
And it took a number of years for that 1 Mbps connection to be used to capacity. After all, this was a 20-fold increase in bandwidth and it took a long time until households began to download enough simultaneous things to use all of that bandwidth. But over time the demand for web broadband kept growing. As cable networks upgraded to DOCSIS 3.0 the web started to get full of video and eventually the 1 Mbps DSL connection felt as bad as dial-up a decade before.
And this is perhaps the major point that the article misses – you can’t just look at today’s needed usage to talk about the best technology. Since 1980 we’ve experienced a doubling of the amount of download speeds needed by the average household every three years. There is no reason to think that growth is stopping, and so any technology that is adequate for a home today is going to feel sluggish in a decade and obsolete in two decades. We’ve now reached that point with older DSL and cable modems that have speeds under 10 Mbps.
The other point made by the article is that there are technology steps between today’s technology and gigabit speeds. There are improved DSL technologies and G.Fast that could get another decade out of embedded copper and could be competitive today.
But it’s obvious that the bigger telcos don’t want to invest in copper. I get the impression that if AT&T found an easy path to walk away from all copper they’d do so in a heartbeat. And none of the big companies have done a good job of maintaining copper and most of it is in miserable shape. So these companies are not going to be investing in G.Fast, although as a fiber-to-the-curb technology it would be a great first step towards modernizing their networks to be all-fiber. CenturyLink, AT&T and others are considering G.Fast as a technology to boost the speeds in large apartment buildings, but none of them are giving any serious consideration of upgrading residential copper plant.
It’s also worth noting that not all companies with fiber bit on the gigabit hype. Verizon always had fast products on their FiOS and had the fastest speed in the industry of 250 Mbps for many years. They only recently decided to finally offer a gigabit product.
And this circles back to the question of whether homes need gigabit speeds. The answer is clearly no, and almost everybody offering a gigabit product will tell you that it’s still largely a marketing gimmick. Almost any home that buys a gigabit would have almost the same experience on a fiber-based 100 Mbps product with low fiber latency.
But there are no reasonable technologies in between telephone copper and fiber. No new overbuilder or telco is going to build a coaxial cable network and so there is no other choice than building fiber. While we might not need gigabit speeds today for most homes, give us a decade or two and most homes will grow into that speed, just as we grew from dial-up to DSL. The gigabit speed marketing is really not much different than the marketing of DSL when it first came out. My conclusion after thinking about this is that we don’t need gigabit speeds, but we do need gigabit capable networks – and that is not hype.