Cloud-native: A revolution that will fundamentally change IT architecture

Cloud-native architecture is the next evolutionary stage of enterprise architectures

IT architecture at companies has historically changed from a mainframe to a service-oriented architecture, separating frontends from backends with an API layer. Another change, which is taking place now, is the shift towards implementing headless solutions, i.e., implementing systems that have only APIs and are expected to integrate with other systems (including digital channels and branch systems). The future architecture will be built on cloud capabilities, on available, ready-made services and tools that we will simply combine must more than implement.

[.infobox]However, by the time the entire enterprise architecture becomes headless, it will be outdated.[.infobox]

This trend will be further reinforced by the emergence of industry clouds. An industry cloud is a ready-made set of systems for a specific vertical (manufacturing, finance, insurance...). The advantage of the industry cloud is not only the availability of business-oriented systems in cloud environment (e.g., insurance-as-a-service), but mainly the fact that these systems and tools are pre-configured and mutually pre-integrated for a given industry vertical. One of the first uses of the industry cloud is the VW Digital Production Platform built on AWS Manufacturing industry cloud technologies.

We're constantly re-programming the wheel. Cloud-native will reverse this trend

It is the easy availability of off-the-shelf solutions that will lead to the very rapid adoption of new technologies. It is significantly easier to use a ready-made system in the cloud than to deal with the project of deploying a new system in an on-premise environment. The adoption of new systems will thus become significantly faster and cheaper. The gap is expected to be so large that 95% of all new solutions will be cloud-native by 2025 (Gartner prediction).

But the shift to cloud-native will also significantly change the way IT is currently developed. Today's agile teams tend to develop everything on their own. This trend has been intensifying for several years and instead of buying ready-made solutions, even complex systems are being developed from scratch. The efficiency of IT development thus decreases because the wheel is constantly being re-programmed. Various frameworks and open-source components are used, but these cannot match the deployment of a more comprehensive business solution. And it is the ease of adoption of cloud-native and SaaS technologies that will reverse this trend. The difference between in-house development and using a ready-made service will be so vast that the latter approach will clearly dominate future development. This will fundamentally change the role of agile teams, their planning and organization.

[.infobox]I expect that once companies gain experience from teams that have been through such a project, and once architects adopt a new way of thinking, the advent of cloud will be literally precipitous.[.infobox]

The key will be to enter the right ecosystem of interconnected tools

The use of pre-integrated business services will bring another big change: by using a specific service, a company will become the participant in a specific ecosystem of other services and tools. These additional business services will be integrated with the first one and therefore will be much easier to use. However, the number of companies that decide to integrate another service (which may better suit their needs) using their own resources instead will be small, as the difference in labour intensity will be huge. Hence, an analogue of vendor-lock will appear, which we can call the ecosystem-lock. I expect that many companies will underestimate the analysis of the entire ecosystem and will become participants in a particular ecosystem by accident, rather than through a carefully managed process orchestrated by their enterprise architects.

Cloud-native adoption is hindered by prejudice and lack of experience

Our experience from cloud-native projects, which we deliver for example to VW or Provident, where the use of cloud-native features is strictly ordered, shows us that such a strict order does not limit the design of the solution at all. On the contrary, it stimulates the search for new options. Based on experience with these projects, it is clear that agile in-house development on an on-premise environment (or in an IaaS environment) cannot compete with the new approach in the future.

Unfortunately, the use of cloud-native and SaaS approaches often runs into prejudices about data security, availability or the opinion of regulators. Not to mention ignorance and little experience with cloud-native principles. The adoption of this trend will therefore not be immediate. Nonetheless, I expect that once companies gain experience from teams that have been through such a project, and once architects adopt a new way of thinking, the advent of cloud will be literally precipitous.