Ambient Computing: How long before Cloud Computing starts to rain?

Ben Hyde writes:

The four horsemen of the digital apocalypse are cpu, storage, communication, and content, aka Moore’s law and his friends mean that costs are always falling, capablities are always increase, and displacement opportunities are always popping up. These are the drivers that keep pulling the rug out from under any industry that’s built on them. The telecom industry, the entertainment industry, the content based industries.

[…]

To take an example the Telecom industry didn’t lower prices for decades. Their network effects made routing around them very hard. But finally the largest customers started to route around them. Much later the industry suffered a vicious restructuring.

Which brings me to today’s question? We have numbers, i.e. communications capacity doubles ever 9 months. So what’s the number for cloud computing and why do EC2’s prices appear to be unconnected the numbers?

I suspect there’s a very simple answer to the stability of cloud-compute service costs over time, and I submitted it as a comment on Ben’s posting: most EC2 users are not paying for compute performance, but instead to have their time freed up from feeding and watering a farm of computers; so much of the charge is in lieu of personal or corporate opportunity cost which is not closely related to Moore’s law – at least in a world where you are running an application where “fast enough” is adequate performance.

I slept on this idea, and I think there’s a codicil to be added; the above is an assumption on the part of those who price compute services, wildly different from the old days where we used to charge for CPU utilisation; you are no longer paying for scarce resource but instead for convenient access to “adequacy”.

“EC2 runs my webserver, it’s very convenient, it copes well…”

…sort of thing; a world of implicit and explicit service level agreements, you want your stuff to run on something that’s virtually a mid-range server because you can’t be arsed to set one up specially…

This model only works until the next round of the cycle of incarnation, so once people are again capable of achieving “adequacy” on their “ambient” computing – cellphones, laptops and always-connected home PCs – then that will make make a big dent in the assumptions which price “cloud computing”, for the masses, at least.

Oh, and Nokia’s making a laptop. Microsoft should be scared; if it’s Linux-based like the N800 / many Netbooks, it’s a huge threat. Give it a few years and the ambient computing capability will be huge.

Comments

3 responses to “Ambient Computing: How long before Cloud Computing starts to rain?”

  1. I love the suggestion that running a data center is like being in the 4H.

    Yes people are buying the value of having somebody else keep those stables clean. Yes they would be willing to pay something close to what it would cost them to run a small family farm. But AMZN is running a factory farm, and it’s costs should be a fraction of mine.

    The degree of pricing power AMZN has is revealed in the difference between the price they charge and their COGs. If they already lack effective competition, and that’s really what I’m suggestion, then your right they can charge for value. That has side effects.

  2. […] but instead to have their time freed up from feeding and watering a farm of computers”.  He followed up in his […]

  3. Ok – this must be a 4H – http://en.wikipedia.org/wiki/4-H – but I don’t understand it. I am in the UK and lack the background on what this analogy means? Once I understand, then I will cogitate and respond. 🙂

    Some sort of co-operative of family farms, maybe?

    A little help, please?

Leave a Reply

Your email address will not be published. Required fields are marked *