The Value-Based Cloud
I got some interesting responses to my last article, "How is Cloud Computing Like Pork Bellies?" both by mail and directly from customers that I shared it with. What the upshot of the comments was that customers seem to want commodity pricing for a service that is completely custom. And they've found some interesting ways to try to reach that point... One potential customer stores 100TB of data in Amazon's S3, but because they're a photo sharing site they can't afford Amazon's "hidden" bandwidth charges for shipping all that data around and S3's high latencies, so they built a datacenter for themselves and put large storage arrays and user-facing computation systems in it to provide the performance they needed while caching their data to keep bandwidth charges to Amazon down. Another of my customers serves a large client in Redmond via our cloud service, and was told by Redmond that it was great that they were using Cloud, but that the client wanted all their data in a separate piece of hardware, which amused me a great deal considering that the client was promoting cloud computing which by definition is based on shared infrastructure. A third prospect likes our service but needs an Apple XServe on the same LAN segment as their cloud resources (to save on bandwidth costs.) Each of these customers was telling me they didn't think of cloud as a commodity because they wanted it customized to their needs.
Then today, I read a number of blog articles by pundits who were making the point that cloud computing is converging on a single worldwide standard which will be available across all cloud providers and make customers happy with the best possible price and the most standards. There's even a startup that is working to this end - Zimory, which aims to offer a cloud computing marketplace. While I'm very excited about open standards, and I'm not even sure if ENKI would exist without them, the Amazonization of computing into a least common denominator doesn't seem to be in line with what most of my customers need to run their businesses. Sure, there are many customers I'll never see because they're only shopping on price, and I applaud these customers because they're building the cloud market. But the bulk of businesses are waiting for cloud computing to be tailorable to their business process and needs before they make the leap. To combine the commodity pricing with this level of tailorability, to my mind the industry needs to make a few leaps before it gets there:
1) Prove the cloud model is financially viable for vendors and customers: OK it works for the web 2.0 sweet spot, but what about the Enterprise or other busineses with custom needs?
2) Agree on standards for deploying to the cloud and managing applications in the cloud, and develop the resulting technology which will reduce prices by stimulating competition but more importantly by reducing both providers' and customers' costs
3) Find ways to capture and address the 5-50% of customer requirements that don't fit into a generic cloud model and are blocking cloud adoption, without breaking the cloud model. Our partner, 3Tera, had this as a founding vision with their virtual private data center paradigm, in which custom hardware is replaced with virtual instances. Yet, the VPDC model still leaves us with some requirements we can't fulfill due to technological limitations.
4) Capture a larger share of the market by having a value proposition that interests business decisionmakers, not just programmers or tech people.
Our CEO, Dave Durkee, presented some of these ideas in a paper he gave last November at the Cloud Expo, on the Value-Based Cloud I'd be interested to hear your comments on it.
After talking about the market in the last few articles, I'm feeling the need to go back to some technology discussions - stay tuned!
TrackBack URI for this entry
