Cloud services are already more relevant than traditional infrastructure services, let’s review the most relevant challenges faced by companies that are going to migrate or are already under an IT architecture model in the cloud.
The migration from infrastructure services to the cloud is already a reality, during the second decade of this century many companies have considered whether it was a good idea to go to the cloud world or continue with their current infrastructure, in that process the TCO costs and benefits were evaluated. Currently, this approach no longer makes sense and to a greater or lesser extent the vast majority of companies have started the migration or are in some cloud computing solution of the main players, Amazon Webservices with 33% of the market, Azure with 18% , Google with 8%, IBM Cloud Services with 6% and Alibaba with 5% lead the top five according to Statista.You will find more infographics at Statista
This process is not being instantaneous but already in 2020 the spending on Cloud services is higher than in the traditional IT infrastructure model. In most cases these projects were pigeonholed within the so-called “Digital Transformation of Companies”, although many clients currently avoid using these terms to avoid what they call an “exorbitant increase in prices” of services under that name.
Analyzing the statistics, we can intuit that the most prevalent current infrastructure architecture model is the hybrid, where customer service channels, whether B2B or B2C, reside or are migrated to cloud environments. On the other hand, the transactional backend together with the data is maintained in a traditional on-premises infrastructure model.
What challenges are companies currently facing in their foray into cloud services:
Cloud services cost optimization
Both companies that were born under native cloud architectures, and those that started the adoption of cloud services in their digital transformation projects a couple of years ago, are currently facing a rationalization of spending on cloud services. What began as an expense of thousands of dollars a month has now increased into tens of thousands of dollars a month without a growth in the same proportion of the supporting business volume. And even if this were the case, it is not reasonable that in order to sell 10% more, the same proportion should be spent on cloud services to meet this new demand.
There are several elements to review in this area, from the inappropriate use of cloud services according to the business need, the scalability and elasticity model of the defined cloud architecture, such as the definition of development models aligned with the good practices defined by the use of digital containers in the cloud, such as DOCKER and KUBERNETES, where applications are housed to optimize resource consumption. Likewise, the review of the contracts and the billing model that each vendor negotiates with the companies is a relevant aspect.
Cloud architecture connectivity
Regardless of the adoption of a microservices architecture, which is highly recommended, it is essential to define the interconnection and interoperability model of cloud applications developed with other systems, as well as integration with transactional backend, legacy access, or to consume data external systems and the exposure of own APIs to third parties.
It is important to properly choose the API Gateway that will be included in the defined cloud architecture, the API manager and its ability to extend its functions, resource consumption, the cost structure for each API call as well as the complement in the management of the APIs make this choice decisive. An opensource API Gateway model will allow for more independence in the future choice of cloud services to hire and will facilitate the adoption of autonomous Multicloud environments.
Evolution of own cloud architecture
Referring again to the “Digital Transformation” projects, there is an implicit message of evolution of the defined cloud architectures, reviewing those initially established in order to take advantage of new standards such as Kubernetes and also to be able to migrate to a complete cloud environment when it is feasible at a technical and financial level, which are the main stoppers of this process.
Evaluating whether it is technically feasible to migrate our transactional backend to the cloud, whether it is a commercial software of the ERP and / or CRM type, or if we are talking about a custom development in two or three layers, involves an evaluation of many factors that in most of the cases is determined by the economic amount involved.
In the case of commercial packages, the evaluation begins by talking with the vendor who will indicate if there is a version in the cloud and who can also indicate an estimated cost of the process.
In the case of a custom development, the evaluation together with the cloud architecture team will define if it is better to try to take it to the cloud as it is, or if it is necessary to make adjustments that make it compatible with the defined cloud architecture. Or if on the contrary, it is better to extract the data and remake the application from scratch. The aspects of licensing costs and development effort will be key to define which alternative is better.