A few years ago I worked for Sony Music and Sony Corporation of America. One of my first challenges as Vice President of Web Operations was to lead a global infrastructure consolidation with the objectives of:
- Lowering Total Cost of Ownership of our entire Internet facing Infrastructure
- Reducing Overall Data Centre Costs
- Reducing the Cost of Power
- Addressing Cooling Issues and Costs
- Reducing Complexity
- Improving Operational Efficiency
- Improving Operational Staff Efficiency
I have recently been asked to re-visit the work I did for Sony and update my thoughts in light of new developments, most significantly the growing acceptance of Cloud Technology.
In recent years, I have developed my thinking beyond Infrastructure consolidation towards scalable target operating models that encompass Data Centre, Data Centre As a Service (DCAsS), Virtualisation and Cloud where the Infrastructure layer supports Business Users consuming Processes that are Globally Common, Regionally Specific or Locally Personal.
Before exploring process consumption further we need to first understand the infrastructure layer and how we have moved from the early days of data centres where we pile them high and squeeze more and more power from the grid to the current trend of consolidation and conservation.
Consolidation is a transition from an unpredictable, inefficient state to one that is sustainable and efficient. (Forrester, 2007)
2007 saw a significant increase towards consolidation. The initial motivations for this move were almost entirely driven by the need to obtain cost savings. Very quickly the motivations moved towards fulfilling the need to improve Business efficiency. Server, Storage and Desk Top virtualisation drove cost savings and the efficiency gains. In terms of Operational IT Staff, it drove a rationalisation, a decrease in costs and a subsequent increase in efficiency. This drive towards consolidation through virtualisation continued for several years to the point where it is now a standard tool in the CIO’s toolkit.
Wikipedia defines Cloud Computing as:
the delivery of computing as a service rather than a product, whereby shared resources, software, and information are provided to computers and other devices as a utility (like the electricity grid) over a network (typically the Internet).
Cloud computing provides computation, software applications, data access, data management and storage resources without requiring cloud users to know the location and other details of the computing infrastructure.
A little thought and it is fairly simple to conclude that once you have commoditised your virtualised infrastructure such that it is securely multi-tenanted the only difference between a virtualised data centre and the Cloud is the commercial model. With Cloud, vendors, making the capital expenditure and sweating their asset on a time based rental scheme.
Moving your operations to the Cloud can have many benefits:
- Speed: With applications that are already cloud-enabled or where Cloud based services already exist, moving your business into the cloud can be very quick and easy.
- Zero Capital Cost. The Cloud model eliminates large initial CAPEX by implementing an OPEX charging model.
- Agility: Cloud services allow you the agility to quickly and efficiently move as your Business moves.
- Scalability: The Cloud offers you the ability scale your infrastructure up or down as your Business need scales. Firms with seasonal businesses or Firms that are project based find this scalability very flexible.
- Low or Zero Maintenance Cost. Since support and maintenance is usually included the need for dedicated IT support staff is removed.
- Reduced End-User Cost. The need for powerful desktop equipment is removed and the life of this equipment is significantly extended because most cloud services are accessed simply via a web browser.
There are, of course, a number of downsides to Cloud Computing:
- Connectivity: Without reliable internet connectivity users will be unable to access services.
- Reliability: Multi-Tenanted infrastructures are inherently less reliable than well-built private systems.
- Latency: Reliance on the Internet and the multi-tenancy profile of Cloud services can lead to poor performance and network latency issues.
- Security: Access via the Internet may be subject to interception and other security risks.
- Ownership: Migration off Cloud infrastructure may cause issues around who owns corporate data and customisations.
- Re-Engineering: Line-of-business applications may require re-engineering or replacement if they are to work in a Cloud infrastructure.
Clearly it is critical that the choice to move infrastructure into the Cloud requires significant planning and due diligence before a final decision is made.
This leads us back to my earlier point about looking beyond the short-term gains, the low hanging fruit, of data centre rationalisation and consolidation through some combination of virtualisation and cloud services.
One has to ask the question, “Why do we have an IT infrastructure at all?” The answer should be fairly obvious:
“An IT infrastructure is there to support the activities of the Business.”
Businesses operate through a given set of processes be it Purchase to Pay, Order to Invoice, Recruit to Terminate or any of the myriad of other processes required to make a business function. Building an understanding of these processes and their relative importance to the functioning of the Business is a critical factor in designing a cost effective, scalable, infrastructure layer that supports Business Users in their efforts to do work and add value. This understanding and a Firm’s progress towards implementing a process centric model is known as the Process Maturity Model.
Larger, global firms have a further complication in as much as their processes can be categorised as:
- Globally Common,
- Regionally Specific or
- Locally Personal.
The simplicity of this model enables the creation of a supporting infrastructure layer for global Firms that extends to the creation of global and regional service centres to allow Business Users to consistently and repeatably consume processes from anywhere on the planet.