The question of Public Cloud versus Private Cloud is one of the most contentious in the field right now.  It generates some very heated debate between evangelists of both camps.  The answer is of course “Your Mileage May Vary”.

 

The most important thing for an organisation making a decision in this area is that the decision is concrete, and transparent in terms of drivers and desired outcomes.  This is often why I feel that the first framework charter on the cloud must be taken by the business, not IT.  (I post some more on this topic here.  Having the business set desired experience and implications drives the truly objective direction.)

 

The first piece to address is what attributes are desired by the word “cloud”.  Definitions abound but all of them will contain at some level the following:

 

Attribute	Private Cloud	Public Cloud
“Delivered on Internet Technologies”	Could be delivered remotely. More often delivered in captive datacentre	Always delivered remotely
“Delivered As A Service”	Service definitions bespoke definable. (Although to achieve benefits, once defined up front they must be then prescriptively locked)	Service definitions defined prescriptively
Elastic Infinite Scale	Reasonably scalable (within the confines of datacentre and capital expenditure limits)	Usually highly (“effectively infinite”) scalable
Pay per Unit (Consumption based pricing)	Unit defined and managed. Not always unit priced. (Depends on maturity of internal IT)	Unit priced and unit cost
Rapid and Agile	Rapid and agile (as agile as IT can make it)	Rapid and agile
Low Cost	Designed correctly, it can lead to much lower cost than existing supply	Consumed appropriately, it can lead to much lower cost than existing supply

 

Choosing the best for your business then becomes a reasonably straight forward decision based a few key questions.  (There are nuances of course, but it is better to choose a path and verify fit than procrastinate.):

 

• Do we as a business believe that this particular service is something unique to us and the provision of which drives a competitive advantage versus our peers in the industry?
• YES: We should make it as agile as possible and deliver internally using private cloud principles.
• NO: We should search for a market solution (likely to exist if it is a very common service), and procure from public cloud, focusing internal IT on other areas.

• Do we believe our internal IT people and systems have the scope, capability and motivation to transform to a private-cloud based delivery paradigm?
• YES: Do not take a half-hearted approach. Design it correctly from foundational principles, and ensure sufficient capital is assigned to enable the transformation [*].
• NO: Redesign the services to consume from a public cloud, or else consider a “classical outsource” of that area to an appropriate outsourcer

• From a business funding and capital asset management point of view, do we have easy access to capital?
• YES: In the long run, an on-premises, fully-depreciated, custom-designed and well-managed private cloud solution will be a very cost effective option with the best Net PresentValue
• NO: Follow the standard decision of “lease versus buy”, and rent it from a public cloud provider.

[*] – (This type of change program may involve reallocation of responsibilities inside the existing IT service delivery organisation as well as injection of fresh blood)

 

Clearly the devil is always in the details in these decision processes, with many variables being involved.  Therefore moving forward requires taking the same approach as solving a mathematical equation.

 

“Progress is impossible until we take a fixed decision on the major variables.  The minor variables will solve themselves, becoming resultant properties of the major decisions.”