brianfending.com

So I heard a humorous suggestion yesterday: "We should have one server per user."

Hm. It doesn't jive with the whole direction of increasingly distributed ("cloud") computing, providing whatever computing resources are needed by a person/application/whatever on demand, even on a tremendously large scale. Of course, this person was referring to a server per user for *one application*, but it's way more fun thinking in the scope of a general one-to-one person-to-machine pairing.

Here's what cloud computing is actually supposed to look like. Note that it begins at the User Interaction Interface and does more than a simple back & forth with a single server (plus mystery resources to the right of it) to handle requests...

There was in InfoWorld article published citing a great example of this scalability over conventional server infrastructure approaches:

Analyst Alistair Croll of BitCurrent said there are specific applications for which grid/cloud computing is perfect. For example, The New York Times recently rented Amazon's grid to create searchable PDFs of newspaper articles going back decades. The Times estimated that the project would have taken 14 years if the Times had used its own server. Amazon did the entire project in one day, for $240.

Seriously. A day. $240. That's insane.

So what of that Matrix-like notion of a dedicated panoply of machines ready to act out orders from any and one? A *virtual* server, a "machine" that starts on command and goes out into a virtual world to return everything we need. This is the space where EVERYBODY wants to play God, from Google to Amazon to Microsoft, and make you feel safe next to your imaginary black box. Computing with a single interface per person - on whatever computer (who cares about the device you use??) - where everything is brokered past that single point of entry... Your Virtual Server. It's the dragon that Killer App folks have been chasing for quite some time now. "Come home to us," they say, while competition has the lot of them reinventing the wheel and drawing you to THEIR single interface.

Once the privacy/security/infrastructure/interoperability concerns (you know, the little stuff...) are worked out - and probably after the monetization is worked out behind closed doors - there is tremendous potential for a single window unto the world that will supplant the browser-based, keyword-search-based Internet we've come to love and waste time on. I wonder, how close are we to that reality?? It really makes me wonder if I should do a Facebook programming challenge in May just because it'll be interesting or instead contribute something to OpenID, develop an S3-based app, work with the Google APIs, or develop a Facebook app.

Or just develop one of those cloud computing infrastructures - I hear it's an easy $240 a day.

• Technology