Tuesday, 31 January 2012

The Race To The Cloud Is On

The cloud is a new Cambrian explosion: When the eco-system changes, a whole new set of animals emerge.

The cloud is a great idea. Devices can overflow their processing and storage requirements to the cloud. This frees them up to become better form factors. Less bulk, more ergonomics. Good news for humans.

And it's opened up a new gold rush: the gold rush to provide your cloud storage.

Let's think about why providers might want to get into that game. The first consideration is that it will be very very expensive to do so. The dominant cloud service providers for Europe (AWS and Azure for now) charge for storage per month, and they're cost model is a reasonable representation of the cost of running their giant server farms.

So providers have to pay every month and the more they store, the more that monthly payment will be. That's very different to buying your own storage outright. If I buy an external hard drive, for instance, I buy it once and it's mine. Effectively, on the cloud, I'm renting that hard drive forever and every time I put more stuff on it, the rent goes up.

Every consumer who's used to the wild and wonderful free Internet won't like that, so to buy their custom, the cloud service providers will have to either pay for it for them, or at least package it up and maybe subsidise it a little to make it buyable.

So why would a provider want to take on that endlessly inflating burden?

Because it gives them leverage. It gives them leverage to provide other profit-making services. Data is king, and the customer is his daddy. Get them into your system, and your system can start making money… somehow.

You might think that providing storage at a loss (let alone the processing that goes with it) sounds like madness. Surely it would be possible to make a tiny tiny profit per customer by marking up cloud storage? At the moment, it doesn't look like it. I'm not sure if they'll last, but Bitcasa are betting the farm on a flat fee of $10 / month for infinite cloud storage. Their model is all about clever compression and de-duplication techniques. If a million people are storing the same copy of Spiderman on the cloud, Bitcasa will store it once and satisfy everyone. But $10 / month?? It just doesn't seem to add up. Dropbox are having a tough time balancing loads of freeloaders with 2+GB against premium customers paying their lowest tier $10 / month for an account limited to 50GB. People who want cloud storage have a LOT of data. And photos, for instance, are original data and don't de-dupe so well.

So who will win?

I can only imagine that the winners of this crazy race will be the people with the largest vested interest. And I think that's not going to be the clever first-mover start-ups like Bitcasa, or even Dropbox, sadly. It'll be the people who have massive scale (deep pockets!) and who want to lock you into a wider infrastructure of services built around the cloud. So that'll be Apple, Microsoft and Google, then.

Good on them. But I am going to put SOOOO much data into their systems that they might regret winning for a few years.

No comments:

Post a Comment