Evaluating the Optimal IT Service Delivery Options

August 8th, 2011Cloud Computing


An interesting article in the New York Times last year, Too Many Choices: A Problem That Can Paralyze, pointed to a problem we can all face at times – having too many choices can be even more challenging than having too few.  As IT service delivery offerings proliferate with new cloud options, this is an issue that IT executives will wrestle with for some time to come.

To help you sort through the options, let’s start with some basic definitions. We tend to group IT service delivery into 5 core models:

  • Internal Data Centers – company-owned and operated traditional IT infrastructure.
  • Internal Private Clouds – leveraging the benefits of automated resource provisioning, lifecycle management, charge-back or view-back, simplified portal-based interface, and a shared resource pool with some level of elasticity, typically owned and operated by the company.
  • Hosted Private Clouds – the benefits of Internal Private Clouds, but shifting the burden of operations and the capital expense to an external cloud host, and typically providing greater elasticity via a larger available pool of compute resources to draw upon.  They also offer greater security than Public Clouds with no multi-tenancy at the physical server level.
  • Public Clouds & Community Clouds – the benefits of Hosted Private Clouds, but with multi-tenancy inherent to the model, thus causing security concerns for certain use cases.  Community Clouds are a variant of Public Clouds and share the same attributes but restrict access to a known set of organizations, a model common in Federal, state and local government.
  • Outsourced Data Centers – transfer of resources and operational responsibility of the IT infrastructure to a third-party outsourcer.

You may be considering variants of a number of these service delivery models, possibly using several Software as a Service (SaaS) applications in addition to sourcing capacity in multiple public, hosted private and/or private clouds.  If so, a structured assessment of your applications, data center services (including outsourced and cloud-based IT) and management is critical to your success.

Applications – Since no single deployment model is right for all workloads, a formalized approach is needed to ensure that workloads are analyzed and correctly placed into the appropriate delivery environments.  To accomplish this, a workload placement decision model should be used to characterize applications and drive static or dynamic placement decisions so workloads can run on the optimal infrastructure.

Data Centers – You will need a consistent approach for building and evaluating data center characteristics, one that meshes with the application modeling above, and that optimizes the data center footprint across cloud and non-cloud delivery models without breaking processes, and which minimizes complexity. We apply our Unisys 8-Track methodology consistently to all delivery models.  This is not a flashback to the ‘70s, but a rigorous approach covering (1) management tools and processes, (2) financial policies, tools and approaches, (3) physical and virtual architecture design, (4) security, (5) storage, (6) network engineering, (7) ITIL / ITSM, and (8) support services.

Management – Finally, you’ll want to take a comprehensive approach to unify the management of these diverse delivery models.  This will not come from a single product or technology, but rather from an integrated framework that addresses five key process areas for all traditional and cloud delivery models: (1) security and risk management, (2) monitoring and event management, (3) financial management, (4) service management, and (5) policy and governance.

If you can successfully apply these principles you will be attaining what we call the vision for a Hybrid Enterprise – one that delivers static and dynamic workload allocation to the most optimal infrastructure, a reduced and optimized data center footprint via cloud technology, and a “single pane of glass” management framework for both cloud and traditional infrastructure.  And, hopefully with this structured approach, you can avoid the ‘analysis paralysis’ that all these great new IT service delivery models may cause.  What are you doing today to address these issues? Please leave a reply below and share your experiences  below.


Tags-   Cloud computing Hybrid enterprise IT service delivery SaaS