Let’s get serious about the cloud!

The cloud is cool. How can it be leveraged to get maximum added value?

We live in the cloud age. Most of the internet services I use run on the cloud. The business applications our customers use are also mostly on the cloud—and not just a private cloud, but often a public cloud too. That’s no surprise. Today, getting started with cloud technologies is really easy. You can set up a cloud account in just a few clicks. And any administrator should be able to transfer a server from an on-premises environment to the cloud.

We already have the first stage in the development of cloud service use behind us. It was characterized by the transfer of IT resources in their on-premises form to the cloud world offered by individual cloud operators. The goal was mainly to simplify the purchase of IT resources and facilitate their management. It addressed problems related to cloud security, secure data storage on the cloud and, last but not least, the compliance of cloud use with legislation and internal regulations. In the first stage, these were also the main obstacles to more widespread use of, particularly, public clouds.

Public cloud operators have consistently zeroed in on these barriers, and they can now provide the necessary information, certifications from the relevant authorities and other documents to dispel doubts in these areas. They can also supply infrastructure in various locations that complies with legislative requirements.

Now, we are ready for the next stage in the development of cloud service use. So, it’s time to seriously consider where to look for more added value in cloud technologies. We also need to start seeing the transition to the cloud as an opportunity to make a qualitative change in the use of IT resources.

The situation is similar to those in other areas where new technologies and workflows are being deployed. Here are four examples.

  • Robotization is not just about replacing human labour with automatons. It is an opportunity to move from cyclical to continuous product processing. This applies to both individual product processing and product innovation.
  • Everyone can certainly agree that digitizing agendas to replace paper documents with electronic versions does not make much sense on its own. However, it is an excellent opportunity to eliminate data duplication and redundant processing steps. Above all, it is an opportunity to improve communication with users and make services more convenient.
  • Communication channel unification—for example, in the banking or telecommunications sector—is not just about creating new all-in-one software. It is an opportunity to introduce an entirely new approach to managing product portfolios with an extremely short time to market.
  • Introducing agile development is not just about renaming roles and redistributing teams. It is a unique opportunity to gain the ability to respond immediately to changes in the external environment, changes in customer requirements and even changes in competitor behaviour.

What opportunities does the transition to cloud services offer? Here are two.

The first opportunity doesn’t need to be explained at length. It is the ability to obtain exact and well-structured information about the costs of IT resources, the profitability of IT and the structure of the IT infrastructure.

In the cloud, there usually isn’t enough information to identify which service was allocated by whom and when. However, a commonly-used simple concept can introduce order to the environment, although we must create and maintain that order ourselves. That concept is tags. We can label each service with a tag. The cloud doesn’t dictate what tags to use—it’s up to us. Many tags are needed for them to be useful, but no one enjoys entering and maintaining them.

Thus, a tag strategy needs to be implemented and required from the beginning. The tag strategy outlines which tags each service must have and which values they can take on. It can be said that the ability to use the cloud effectively is directly proportional to the quality of the tag strategy.

The most important tags are those for classifying resources within an organization—organizational units, products, business processes, etc. To put it simply, the structure of the tags must correspond to the structure of the organization’s internal accounting. This part of the tag strategy has to be handled by the finance department or controlling. IT and project perspectives are also necessary but less important. From a maintenance perspective, tags are useful for identifying a device’s business and technology guarantor. If the resource is being created as part of a project, it is good to add tags with information about the project too.

The infrastructure as code approach can help ensure that all tags are completed. Tools that support this approach allow you to maintain all the necessary tags for any resource automatically.

Setting up a tag strategy is not just an opportunity to expose the structure and costs of cloud services—it’s a necessity. Experience has shown that it is often cheaper to create a new system than to shut down an existing one without compromising business processes. In the cloud, where the number of resources created is in the hundreds or thousands rather than in units, interfering with a misidentified service poses a significant risk that no one wants to take, regardless of the costs of running it.

The second opportunity to gain more added value by moving to the cloud is a bit more complicated to explain. It is the real use of the ability to scale cloud services according to user needs. This means focusing on performance management and adopting an approach called “workload isolation”.

In the on-premises world, the main limitation is due to the power of the available hardware. As a result, a lot of effort and money is spent on optimizing processes so that processing takes place within a user-acceptable timeframe with the existing computing power. The trade-off between functionality and processing speed is constantly being addressed. Implementation of the required functionality is often delayed because it would overburden the existing system or prolong its response time unbearably. On the other hand, if there is enough power, the complexity of unnecessary functionality or service applications that are not directly needed from the user’s point of view is not addressed because the required computing power simply exists.

In the cloud, it is completely different. There are basically no performance limitations. If a functionality needs power to process at the required time, it is no problem to allocate the power. Moreover, this means that power only needs to be allocated at computation time—if power is not required for a particular business function, it is not allocated.

This changes the whole paradigm. The on-premises philosophy says that the installed power should be used to the fullest. On the other hand, the cloud philosophy says, “Use whatever you need to satisfy users, but think twice about what is really necessary and when. And most importantly, get rid of all the ballast.” This is especially true of systems and functions that are not directly related to business processes or do not have a specific customer.

Such a paradigm shift will undoubtedly lead to a change in architecture. Specifically, architectures that separate stored data from computing capacity (workload isolation) are emerging. The data has to exist long term, regardless of when it will be reprocessed. It must exist even when no computing power is required, such as at night. Therefore, the means for storing them must not be tied to the means for processing them. For the calculations, computer systems must be used that allow the performance to be scaled to an extreme extent. Additionally, it is good to focus on systems that allocate and deallocate computing power as quickly as possible, ideally instantaneously.

The cloud, just like any other technology, has its advantages and disadvantages. If we want to use cloud technology, we should do it in a way that brings users the greatest possible added value. At Profinit, we have been actively engaging with this topic for five years, and we are happy to share our experiences.


Author: Ondřej Zýka

Data and information are assets that need to be properly cared for. Ondřej Zýka has been promoting this idea for over fifteen years as Information Management Principal Consultant at Profinit and has been guided by it in projects involving data warehouse development, data quality improvement, metadata management and master data management deployment. It is also one of his lecture topics at universities, where he lectures on database systems.