A blog on the ipt event «cloud-native development» on March 31, 2021.
Two companies, two exponents, two opinions - that's nothing new. But in the discussion on March 31, 2021, at our «Impulse Live» event on the topic of cloud-native development with Marco Peyer (Swiss Re) and David Ulrich (AXA), more similarities than differences emerged. Being successful with cloud-native seems to be based on a few, but important principles.
Developing applications cloud-native simply means developing applications in the public cloud and exploiting the full potential of the underlying hyperscale platform there. Instead of worrying about VM's, firewalls, frameworks, CI/CD pipelines or monitoring solutions, the focus is completely on the business functionality of the software solution. All other functional and non-functional requirements are abstracted by the underlying cloud services. Be it the detection of damage on images, the identification of the customer (IAM) or even the provision of the complex infrastructure. This allows companies to quickly and accurately adapt to changing market needs. Cloud-native has emerged as a success factor for data-driven digital software solutions. However, there are risks that need to be considered in cloud-native development; more on this in the coming sections. If you want to learn more about cloud-native, read the blog speed in development thanks to cloud-native.
Just a few years ago, only the big players on the market could afford an innovative and modern IT infrastructure. But with the triumph of the public cloud, this market advantage is disappearing. Thanks to hyperscalers such as Microsoft Azure, AWS or Google, even an SME or a start-up can obtain a top infrastructure without any investment. The two guest speakers Marco Peyer and David Ulrich emphasized that they would have no chance of operating an IT infrastructure that even comes close to the innovation potential and efficiency offered by the public cloud. In addition, the hyperscalers promise to contribute to a reduction in CO2 emissions, as they make more efficient use of the existing IT infrastructure.
It was also impressive to see how the corporate strategy and vision of a company changes massively in the wake of digitalization and the public cloud. The development of opportunities in terms of growth, differentiation and customer loyalty was clearly demonstrated - based on technological progress through cloud-native.
Swiss Re aims to be 100% in the public cloud by the end of 2022. AXA is aiming for this goal by 2025. While Swiss Re primarily focuses on one hyperscaler, AXA relies on two hyperscalers with different use cases.
Dealing with the regulators was a central topic in the subsequent discussion. While the path to the public cloud appeared simple on the slides, the road is rocky and not straightforward. Different requests to speak confirmed the impression that the maturity of the hyperscalers with the topic, as well as the changing requirements of the regulator, turn out to be the greatest challenges.
Once this is achieved, the question of lock-in or business continuity arises. While AXA has agreed on a common process platform to eliminate the lock-in regarding hyperscalers, Swiss Re relies on the simplicity of a hyperscaler to fully benefit from the advantages. In both AXA's and Swiss Re's strategies, it was clear from the discussion - there is always a lock-in. The only question is where to put the lock-in and how to mitigate the lock-in with pragmatic architectural means.
Swiss Re and AXA have already made their way into the public cloud and implemented well-known cloud-native projects. But the question is, what was the secret recipe for achieving this in a relatively short time? In both cases, a central program was responsible for the vision, implementation and (financial) support of the projects. Both emphasized that this had been the central pivot for success. After all, the path to the public cloud is not free and requires cultural change that must be supported from the very top. Many workflows, processes and skills are no longer in demand or are undergoing massive change. Employees must be actively involved in this journey.
Based on the discussions and insights into the public cloud strategies of AXA and Swiss Re, 5 tips for successful cloud-native development can be derived:
Ensure that the development team takes responsibility. The team should play a central role in pragmatically balancing the aspects of technology, cost and the needs of the business. And it must be responsible for ensuring that logical abstraction is incorporated into the architecture. Logical abstraction ensures that individual components can be replaced quickly and efficiently. In the technical language these responsibilities lead to «software craftsmanship». This is understood as the pride and responsibility that everyone in the development team takes in delivering a successful project.
Ensure that the cloud-native solution is guided by the criteria of flexibility, iterative development with a clear vision in mind, maximized use of managed services, multi-channel capabilities, and focus on business functionality. This will ensure that the development team is guided by a vision and can make the right technical decisions.
Enforce a clear and transparent cost structure. Involve the development team in the responsibility. This helps the organization to keep the costs of the public cloud under control. As an additional layer, governance mechanisms can be used to restrict the use of particularly high-cost services or to require the use of lower-cost alternatives.
Accompany the path to the cloud with a program or transformation management. Aspects such as skills management, cultural change, roadmap, financing, and technology management must be covered. In this way, you set a clear signal for the employees where the journey is leading. And everyone pulls in the same direction, or unpleasant surprises or inefficiencies are minimized.
Take a pragmatic approach to the much-talked-about lock-in. There is a lock-in in every case. In terms of calculations, it is rarely worth actively circumventing supposed lock-ins (e.g., with multi-cloud, abstraction layer). It is much more important to be able to implement one's own goals efficiently with the selected technologies or hyperscalers. Risks can be actively managed with conscious architectural decisions.