John Willis, Cloud Consultant
You’ve built up quite a bit of content blogging about the cloud at your site johnmwillis.com in just a short time.
I’ve been doing IT work for 30 years, including a lot with IBM Tivoli, but I’ve been focused on the cloud stuff for two years. I was lucky to meet some of the real pioneers in this space who gave me a kind of baptism, and since I’d been working with large corporations all my life, I completely got the idea. I started interviewing those same people when just a few of us were out there having this conversation.
Do you see the cloud as more than another evolutionary step in IT?
As I picked it up from others, I also came to think that the cloud is kind of a Cambrian explosion that may mark this time in IT history as a spectacular moment. You could say it started with computers and IBM or PCs or the Internet, but the convergence of massively scalable commodity computers, open Internet protocols and our understanding of it looks different than everything I’ve seen in the last 30 years. There’s plenty of hype, but the stories are extremely real and you can’t ignore them.
How have you seen the cloud change a corporate approach to business?
Just look at Google, Twitter or Flickr, these are unbelievable growth stories. Twitter may be the fastest data growth story in the history of mankind. Regular corporations need to start thinking that way. There’s a ship going left and a ship going right, and which one do you want to be on?
Apart from Internet models, how do you make this message resonate with clients?
Those companies are real businesses too. Barnes & Noble and Borders thought they were doing fine until Amazon.com came along. I use the animoto.com story with clients. Animoto was four guys who had a good idea for turning pictures into music videos, but then you explain that they went from 50 servers to 4,000 servers and from 35,000 users to 750,000 users in almost no time. If I could talk to [CEO] Jamie Dimon at JPMorgan Chase and say, yeah, that’s an Internet thing, but would you like to go from 35,000 users to 750,000 users in a week? Could you do that? Sure, he leads a worldwide banking infrastructure and animoto is just a niche business. But these days I’m not sure there aren’t two guys sitting around somewhere figuring out how to do banking with almost zero startup cost and zero touch.
Is infinite computing scale the main attraction of the cloud you are pitching?
Yes and no. Eli Lilly is a great story. Before they switched to cloud computing for their research they would literally wait eight weeks to get a server provisioned. Now they wait eight minutes. That’s not infinite thinking, but it’s changing the concept of time at their company. It’s not only weeks to minutes, they went from thinking they wouldn’t even test a question because of the effort, time and cost to saying they could test anything with that frame of mind. Now they can spin up 1,200 servers by hitting a button. The fact that should attract everybody’s attention is that your average Joe can now engage computing power that a few years ago was only available at great cost at maybe five or 10 places in the world.
You’re very interested in open source software approaches to cloud infrastructure.
I really believe in people building abstractions on what other people have built and adding value. If Google created the base for what Amazon came to do, then Eucalyptus, which was a college project, turned out to be an abstraction for private clouds. In the same way Amazon changed the world for people willing to work outside the enterprise firewall, Eucalyptus has built an abstraction layer that will allow the enterprise some day to run their own Amazon in their own infrastructure. You need to look and learn about this stuff. Some people have heard about Hadoop, but they don’t know that people who have built abstractions have turned it from rocket science into something a lot of people can use easily. At a conference I met a gentleman from Fox News who is using Hadoop to figure out advertising placement from analyzing MySpace data. I thought that was pretty cool.
Isn’t the conversation between an enterprise architect and businessperson going to revolve around established vendors?
If you go to any big company you will find people using Amazon Web Services. I see enterprises trying to put their fingers in the dike and get a handle on their internal customers. They will try to get some level of cohesiveness - between some gunslinger experimenting with the cloud and corporate trying to align these activities. I know that world, and I think enterprise guys have a lot of things working against them. Who are they going to listen to about the cloud? They’re not listening to the open source or Eucalyptus guys, they’re hearing IBM or HP or CSC, whose vision of a cloud is really virtualization and consolidation. I don’t think they ever really get what a cloud might look like, or they don’t want to hear it. The big vendor story is how to get the customer into their own server-centric cloud world. To understand the difference between the Eucalyptus cloud and VMware’s cloud, you have to have a better understanding of how you can do something yourself that would have been crazy five years ago. You don’t have to be JPMorgan Chase or Eli Lilly to do large scale data processing.