I’m looking forward to Synergy in a couple of weeks. Err, one week…

I wanted to provide a glimpse into the Data Center & Cloud agenda at Synergy, by alluding to a key theme that I’ve seen recurring in our discussions with enterprise and cloud vendors alike over the last year: a transformation from what one could think of as a first-generation Infrastructure as a Service (IaaS) model that provided a relatively simple virtual private server abstraction that is best suited to next-gen web based apps, to a richer infrastructure model that aims to serve today’s enterprise workloads.

The increased interest in moving enterprise applications to the cloud has come as the economic and business benefits of cloud computing have become better understood by CIOs, and enterprise users have begun to seriously consider the use of cloud-based services to host their workloads.

Frequently I hear that today’s clouds aren’t “enterprise class”, when in general that is not true. Certainly the degree of automation, the focus on security and correctness of operation in today’s clouds equals or exceeds that of any enterprise. The business model of the cloud depends on it. Perhaps what’s missing is a way to prove this assertion, but I’m sure that in the near future you will find plenty of data about availability and service guarantees for any provider. The nagging questions will include security and compliance, and concerns about data loss, location and so on. I don’t specifically want to address these issues because they have been well covered in the press and by analysts. I do want to point to a very significant trend toward developing a richer infrastructure platform that is much more useful to todays enterprise IT administrators, as they contemplate moving additional core business apps to the cloud. I call this the trend toward the “enterprise ready” cloud. In my definition:

  • Enterprise Class refers to attributes of a capability set
    • For example Amazon EC2 offers an Enterprise Class infrastructure service
    • So does Carpathia, and other vendors we will talk about at Synergy
  • Critically, their IaaS offerings are quite different, though both are enterprise class. In what ways? Below I list a few
    • The kinds of infrastructure abstractions they offer: for example using constructs similar to those used by IT to build and manage today’s applications in the corporate data center.
    • Their business model, billing methods and customer support models: for example long term contracts with both dedicated resources and optional rented facilities, billed per month, with caps
    • The richness of their administrative and operations management interfaces: for example the support of granular role based access control, AAA, logging and ITIL like workflows
    • Their suitability for different enterprise workloads (for example, ephemeral compute from EC2 versus dedicated hosting with optional elasticity from Carpathia)
    • Their focus on regulatory compliance, which demands key infrastructural properties, starting with rigorous multi-tenancy, but extending up the stack
    • The richness of their network service abstractions: for example offering their tenants an ability to manipulate the virtual network, manage it, inspect traffic, and interpose typical enterprise “bump in the wire” services

There are many more, but these make a good start. Some of my views on this topic are available from a panel this week at Interop, where I found a fellow traveller in Randy Rowland of Terremark, who is building just such a cloud.

At Synergy we aim to make the concepts of cloud concrete, and demystify the topic so that you can know what is real, and what is still … err… cloudy. Come to Synergy to hear about our practical, concrete success stories, and to learn what you can do today, and what we think you can do over time as vendor solutions mature. Finally, a few comments on competitive offerings are mentioned here, in a quick video blog I threw together in the early morning before the Interop panel.