zAgile exhibited at the SDForum called "Shaping the New Age of Application Development," in Santa Clara, CA this past Friday and Saturday.
A couple of cloud thought leaders provided keynotes each day:
- James Staten, Forrester
- David Chappell, David Chappell & Associates
Unfortunately I missed James Staten's presentation, but I heard rave reviews,and I enjoyed looking over his slides. I did attend David Chappell's presentation and found his presentation to be quite an interesting way to view the various options for cloud computing today.
I tweeted some of my thoughts, and they crystalized more over the next couple days.
With all the talk of cloud, I felt a bit frustrated. On the one hand, it seems that everyone agrees that for ISVs, using the cloud is a slam dunk because their hardware requirements can scale up and down at will. But apparently it's not obvious yet on how the cloud is going to make money while serving the enterprise. It seems that everyone weighed the benefits of PaaS vs IaaS, what kind of computing do you require, how big do you need to scale, etc.
I am not saying I have it figured out, but here are my thoughts... and I wonder if it's a lot simpler than all that.
My basic feeling is that to understand where the enterprise (ie, “the money”) will use the cloud, it is first important to recognize the strengths of the cloud and how it can save money for enterprises. Simply put, the cloud is great for “instantaneous” and “elastic demands” of customers. “Scaling” is the wrong word here in my view as I pointed out in one of the break-out sessions because if infers “scale-up” while diminishing the importance of “scale down” instantaneously. (Someone tweeted nicely that Sun brought us scaling, and Amazon brought us elasticity). These elastic and instantaneous services are valuable in this context and could be charged at a premium in terms of a per minute or hour basis but might be more expensive over the course of a year. No matter, I will pay a premium if I am only paying for a few weeks for something critical and urgent.
Therefore, I propose to break the enterprise application types into 3 types:
a) Permanent: traditional, mission-critical applications
b) Temporary: sandbox applications, prototypes, temporary projects, a holiday campaign, etc.
c) Unpredictable: we are launching this new app and have no idea where we will attract 100 or 10,000 users and over what course of time.
I propose that the Cloud vendors have the immediate opportunity in b) and c), and this is where a proving ground is offered to the enterprise. Then, over time, they will feel comfortable with the benefits of the cloud, as well as the cost analysis saving for taking on “a)” eventually. I feel like people are trying to figure out a), because that’s how enterprises traditionally think – because the flexibility that the cloud offers has never been an option until now.
Finally, while I can agree that all cloud vendors may jump on the “reservation” pricing bandwagon because “the other guys are doing it,” I do not believe that customers will buy-into it in a big way until b) and c) are proven out with the early majority of enterprises.
