Keeping control of your cloud spendings

In the hybrid multi-cloud era, all resources and services are constantly accessible, facilitating the development. There is no need to think upfront about the remaining capacity of storage, compute power... However, the possibility to create as many resources goes hand in hand with the ability to generate considerable costs. Daniel Sahler, Head of Multi-Cloud Services at Fujitsu Luxembourg highlights some points of attention.

Finding balance between flexibility and costs

A first possibility is the use of resource reservations. The standard paying model is called “pay-as-you-go". You pay what you use. When you stop using it, you stop paying for it. This model provides the highest flexibility but it also comes with a highest price tier.

When you do not need flexibility and know in advance which application, which workload, which server to use for the next year, or even the next three years, you can leverage resource reservations or Reserved Instances. With the latter, you will reserve a specific amount of resources for a determined duration. This brings up to 72% on cost saving compared to “pay-as-you-go". If you now also use Windows Server and/or MS SQL Server which you host on Azure, you can also leverage Azure Hybrid Benefits, which adds again more savings, in sum up to 80% in total (if combined with three-year reservation).

The location variable

Deploying resources in distinct locations, also called regions, have an additional impact on your costs. Indeed, not all services cost the same throughout the regions. There are always slight differences. If it is applicable, paying close attention to the point could become beneficial.

Finding the right sizing

Another angle to look at is the aspect of sizing, a question of efficiency. Only if you right-size your environment, you can be sure that you, on one side, only pay for what you need and on the other side, get the performance you need.

But what does right-sizing mean? You need to monitor closely your servers and applications and take continuous actions based on your investigations. Right-sizing is not an action to perform once the application is deployed. Right-sizing is a process: you have to constantly monitor and adapt. In most cases, the initial sizing of the resources is not adequate from the beginning or changes over time. Some resources can be deployed just when they are needed, scale up or down, in and out. An adequate automation enable also cost management.

Getting the right tools

To achieve this continuous sizing adaptation, the right tooling is needed. You can leverage the tools provided by the platform, which differs from provider to provider. These tools are adequate but limit your view on the given scope.

Now, you might find yourself in a situation where you have to deal with one or several on-premises datacenters, all connected to again one or several cloud providers. How do you keep control of the costs then?

Additional tools provide a clear and broad view of this constant resource and cost management that enable cost optimization. For instance, Turbonomic (IBM) is a software that connects to your environments and leverages the data which is already exposed by your applications and infrastructure. Thanks to this data, real-time analytics are performed. They can generate actions in real-time, whether planned, as part of an approval or of a DevOps workflow.


Daniel Sahler, Head of Multi-Cloud Services at Fujitsu Luxembourg, and his team of experienced Cloud Experts are always open to discuss and to draw your way towards a robust and highly efficient hybrid Multi-Cloud architecture.

Find the original article in French in the LG magazine by clicking here