A Walk in the Clouds

Public clouds are now just a part of the landscape.  For those that attended last year’s NonStop Bootcamp, you know that NonStop is creating a model for public cloud enablement.  By changing the underlying fabric to an Ethernet based fabric and by creating a Hypervisor layer that would support many/any hypervisors then the major issues for launching NonStop in a public cloud are resolved.  I imagine this to continue moving forward and look forward to the updates that will be provided in this year’s bootcamp.

That being said public cloud should be considered carefully and with a clear vision of what should be measurable benefits.

There has been some confusion around what is sometimes referred to as the “Everything to cloud” strategy.  That is not a strategy but a wistful direction.  A strategy needs to align with the business and any IT strategy needs to provide tangible business benefits.  Some discussed benefits to public cloud are that the company now uses opex versus capex.  An operational expense is often easier to cost justify and the costs sometimes drop into limits established for business units but there is also no long-term write-off such as there might be with a capex purchase.

Some work needs to be done concerning this shift and whether it’s actuals are in the best interest of the company or it’s simply a way of achieving shadow IT.

One of the more compelling cases is that the business is just charged for what they use.  Anyone who’s played with public cloud knows that if you forget to shut everything down you are charged for things you don’t use.  That’s why an extensive implementation plan must proceed any real strategic migration to public cloud.  Additionally, that implementation plan should include an exit strategy should the presumed benefits not materialize.  The implementation plan should contain what benefits to measure and ranges of acceptable returns. I’ve attended presentations which state providers (Amazon, Google, Microsoft) are better/more experienced than the IT/Security people in the company’s datacenters.  This really needs to be measured on a case-by-case basis but one could argue that the datacenter employees know much more about the specific company requirements, strategies and direction than the providers.

And while the statement that the providers people are more experienced does not mean you get exclusive access to them.  They are a shared resource and that should be taken into account.

A public provider can ramp up and ramp down resources immediately based on demand.  That is a main benefit of a cloud provider, and it fits the Agile development model very well.  The biggest mistake in migration to a public cloud provider is assuming you automatically get all the benefits.  A prerequisite to a successful public cloud migration is to list the benefits you believe you will get.  Determine a way to measure the benefits and have an exit strategy if they are not achieved.

Gartner used to talk about bimodal IT, which are the two main divisions of IT.  There is the reliable operating model with its trusted and true applications usually associated with run books and best practices that have developed around these mission-critical systems. The new part of IT revolves around Agile design and fast-fail business ideas that can quickly be confirmed or eliminated. These are competing IT strategies to a certain extent but the idea is to provide the very best, most efficient and most cost-effective processing to the business allowing it to compete.

This most likely is a Bimodal approach using datacenter and cloud computing.  With some due diligence it can work.  So if you are considering a migration of your NonStop environment to a public cloud when it becomes available be sure to do the proper diligence to ensure that there are benefits, and just as important, that when you enter the cloud, that they are achieved.