Wednesday, January 03, 2007

Commoditised web - how much? what next?

A few weeks ago the company I run, Fotango released our utility billing model onto Zimki which is our commoditised web operating environment.

We are going to blog about this in detail on the Zimki Blog and also the Fotango Blog but I'll provide some information on here for those interested in the ideas behind commoditisation of IT and the stuff I talked about back at EuroFoo and EuroOscon.

As we run our own sites on Zimki, I can now reveal the cost of running our websites in terms of JavaScript Ops, Bandwidth, Storage etc. For simplicity our billing system converts all this to tokens (think electricity kw/h) which can be bought in batches of 100,000 for £1.

So let me be clear, this is a scaleable and robust environment where you can build and release entire applications online without ever needing to set-up a server, arrange a hosting environment, install and configure a database, install a web server etc. There is no initial capital investment for your service, as it is utility basis.

All you need is a browser and knowledge of JavaScript (which runs both front and back-end with an extended persistence layer). You can build an entire business online with just a browser.

So, how much does it cost to run your own online site? Well, let's get a sense of perspective first.

Six years ago when we first created Fotango we spent a large amount on capital to install web servers, set up hosting etc. Setting up a web site could easily run into tens of housands of pounds, if not more.

So today, well :-

  • The Fotango site costs 63 pence per day, or about £19 per month
  • The Zimki Blog costs 19 pence per day or about £6 per month

Of course, we are not charging at the moment - i.e. it's free to use.

The entire business is built around the capabilities of a scaleable cluster with some nifty financial models based on utility, which charge on the basis of how much of the cloud you consume.

For a company web site and a blog - £25 per month seems reasonable - crikes I wished we had this a long time ago.

But that's progress.

However, when we open source the system (planning to do so, later this year) and release the grid elements, our customers will be able to switch their applications and data between different providers at the press of a button. (why? well price and QoS will very between providers - we intend to provide that information along with a number of market capabilities - think energy trading, and I'll comment on this in a moment.)

Or our customer could host Zimki themselves.

Or they could use multiple Zimki providers.

Or (and this is my favourite bit) our customer could even sell spare capacity in their own data centres back into the Zimki grid.

That's the plan - and slowly we are getting there! As for why open source - well see my earlier post about this.

Will I be shifting this blog to Zimki - of course! I'm just waiting for my friend James Duncan to finish writing his Zimki blog application, and then hopefully I'll be able to get the code from him and create my own! (several beers may well be in order)

So why the comment about energy trading?

Well you can build an application in Zimki (either in a realm or using multiple realms) which can have different end-users. So for example, I could build a CRM system for different companies using either seperate sandboxed realms or keeping the data separate with program logic in Zimki.

This means I could build a SaaS application on Zimki (we are intending to release an example billing module for our Zimki customers in order to help them do this, but that's much later).

That means I can create a SaaS application, without investing in hardware, setup etc etc or being concerned about scaleability and such issues.

The end-users of my SaaS application would then pay me (if it's any good!) for it's use and I would just pay the Zimki grid for consumption of resource.

This reduces my risk and capital investment when creating a new business, enables me to expand easily with the business and makes it easier for me to get such services out there.

It also means that once my SaaS application is successful, then shifting from one provider of the Zimki grid to another could significantly reduce costs.

That creates competition and enables price to be balanced against QoS.

The world of commoditised IT is rapidly approaching, which is good because once all established we can get onto the business of commoditising something else - like the manufacturing process (think digital fabrication, most likely with inkjet, giving compositional and geometric freedom)

I look forward to a day when I'll get to read a book from Nicholas Carr on "Does it MATTER - Digital Fabrication and the corrosion of competitive advantage" as we all merrily inkjet print huge numbers of physical goods (as dissussed back at Euro Foo '04).

That will cause even more nashing of teeth than this current round of commoditisation, but then biological manufacture (Drew Endy et al) isn't that far into the future either.

Still, from all accounts, the printing of object and electronics is happily jogging along. (Thanks to James for spotting the post)

Post a Comment