The Importance of Capacity Management Planning in the Cloud era

“Capacity management is the most underestimated problem of cloud computing.”   Morgan Stanley strategy director Evangelos Kotsovinos.

I recently came across the above quote from Evangelos. He hits the nail on the head for most businesses who have taken the step to move to the cloud. Capacity Management is fundamental and highly underestimated, but why? Hopefully I can shed some light on this and help others from shying away from this discipline and change the mindset of what you need to think about before moving to the cloud.

Reduce cost of IT Legacy operations

The saying in recent years has been “move to the cloud and all will be cheaper”. This has been based on the premise that large scale cloud infrastructures utilise commoditised, high density computing infrastructure on such massive scales that the economies of scale simply dictate lower operating costs. However, we are seeing almost every IT Enterprise struggling to keep public and private cloud costs under control.

Cost reduction in the cloud is all about Command and Control. Command of knowing what is being used, where and how, and having that Control over the ever-expanding computing footprint. The critical capabilities for Command and Control of IaaS-based services include:

  • Knowledge of what is there today
  • Knowledge of how much is being used today
  • Pre-migration full costings analysis across multiple cloud providers for specific services
  • Continued measurement of utilisation and costs post migration
  • Heavy dose of automation to help drive control and reduce manual intervention

Given these basic principles, here are my 5 areas to ensure you get the most out of your cloud strategy and what you need to carefully think about and plan for when moving to the cloud.

  1. Full understanding of what is there today

This is the most fundamental step of all, and yet the one that most organisations struggle with. It isn’t just about knowing what IT assets you have, but how they are related to deliver business services consumed by the end-users. It is also about understanding how much resources are being used across the entire set of services. How the resource consumptions fluctuate based on business events. What the service bottlenecks are. And what resources are underutilised.

To address this, you need a combination of mature service design, service transition and capacity management processes coupled with automated cloud-enabled discovery, application service mapping and capacity planning tooling, all working hand-in-hand.

  1. Knowledge of what is being used and expenditure

Many organisations build their cloud migration business case using a set of assumptions about their current data centre costs. Unfortunately, these costs are not granular enough and typically exclude finer infrastructure details such as power, cooling and networks. In the old world that is not a problem as most IT departments do not cross charge or show costs back to the business. However, cloud service providers account for every penny and they issue bills for resources otherwise not fully accounted for.

Capacity planners have a significant input into the cost modelling of services.

  1. Have a clear view on cloud cost options before migration

Clearly the ability to model costs and build a business case for migration to the cloud needs to consider costs across multiple cloud providers so you can choose the most appropriate pricing plan to run the service on. Knowing what you have, how they are being utilised and how much you are paying for them prior to migration are the building blocks to understanding costs post migration.

The more automated this process is, the more accurate and effective the outcomes will be. This is a task that sits firmly within capacity management discipline.

  1. Assign Cloud Tags correctly

Cloud tags are a labelling mechanism provided by the cloud providers to help carry out functions such as automation, manage security and track assets and costs. They are designated to assets and services and are a fundamental enabling capability for tracking service ownerships and associated costs. Ensuring your tags are consistent across your services and cloud providers is just as important as knowing how much resources you are using.

Cloud tags are discoverable through cloud APIs and are integrated into some capacity planning and discovery systems.

I drill down into more detail on what best practice typically looks like and how to apply a consistent tagging plan in my other blog 7 steps for managing multi-cloud expenses.

  1. Continue monitoring expenses using baseline costs

So, we know what we have, how they are being used, and how much they are costing us. The business case is built, and the right service provider is chosen. We now need to continue monitoring how much resources are being used, and whether the cost models we built prior to migration are still on track.  This is a critical function that should be owned and carried out by Capacity Managers driving continuous service improvement and minimising costs.

In summary, the old ways of standing up and operating services in data centre are no longer viable for the cloud. Whereas in the past Capacity Management could take a backseat, it is now becoming a critical discipline for managing and controlling services running the cloud; supporting key functions in migrating services to the cloud and providing the visibility and control mechanism for operating IT services efficiently once in the cloud.

Keyvan Shirnia
Head of Managed Services

By Daniel Swann