Apr 28

Warp Drive would be cool!

If I said:

“The rise of the warp driveis more than just another speed shift that gets people excited. It will undoubtedly transform the space transportation industry, but it will also profoundly change the way people work and companies operate”

You would almost certainly say “Well, duh, but warp drive doesn’t exist”

And that is my problem with this statement:

“The rise of the cloud is more than just another platform shift that gets geeks excited. It will undoubtedly transform the IT industry, but it will also profoundly change the way people work and companies operate”

—The Economist, “Let it Rise,” 10/23/08  (hat tip: Mike Patrick)


I’ve spent the last 6 months getting very familiar with cloud computing, and I struggle with this – not that I think it is wrong, but that I think it is farther away than many think.   Not as far away as warp drives, but still, not something we’re going to see change the industry in the next 2-3 years – maybe 8.   (people generally overestimate what can be done in one year, and underestimate what can be done in 10).  

Right now, CC is a concept looking for a business practice.  Amazon EC2 is nice, but you wouldn’t want to build a typical web busines on it, because your costs would grow linearly with demand, and that’s not where you want to be.   Google Apps are nice, but they seem only modestly easier and cheaper than your basic virtual server hosting package 

(for example, I pay a whopping $5/month for a server with lots of resources for my blogs and Bellygraph)

Cloud Computing has to thread the needle of “wow, I can build a business on that”, and I haven’t see it do that  yet.   I could be blind to evidence, I could be caught up in my outdated web 2.0 dogma, but, on the other hand, I want CC to succeed.  I just haven’t seen the evidence.


Most of the common problems I see thrown around as “CC killer apps” don’t work for me:
  • Turn on a server as necessary: This is already getting pretty easy with virtual servers.  Also, if your app isn’t designed to take advantage of N servers, having access to N servers doesn’t help. 
  • Scaling applications: we already have Akamai and other CDNs to scale delivery of images and other big objects, and they already move the data close to the user.  

Here’s the kinds of things that I think would allow CC to take off:

  • automatically scaling web applications, at minimal incremental cost – this is CPU, filesystem, memory and database scaling – literally you write your application, and you don’t have to worry about where the data is, or the CPUs or anything – everything parallelizes automatically.
  • flash mob insurance – a way to hook in to a CC system to let your website scale automatically when needed, and then dial back down when not.   Again, you need your app to be designed for this.
  • Low-cost massively parallel analysis – this is something we can do today with Amazon, but it is cost-prohibitive.  Plus there’s not enough demand out there to build an industry on this
  • Minimal latency applications (such as multi-user games) that automatically migrate to be closer to the customers
I think there’s a potentially interesting opportutnity for online gaming using just the solutions we have today – if you make the game’s “real estate” follow the network topology (that is, if you’re using the Level3 network in North Atlanta, you’ll be able to interact with other people who are using the same Level3 network).   That would give the users a heightened experience (reduced latency), and scale well with demand, but it would require a very different kind of game.
But I have yet to see the game changer.  I wish I could, because I want to believe 🙂

Leave a Reply

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

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>