Has anyone recently implemented platform engineering? Could you share the pros and cons that your team is experiencing?
Sort by:
Senior Director - Solution Architecture in Software17 hours ago
You will find different pro and con bases on whether (1) your platform is for use in organisation or (2) to deliver customer products faster (like re-usable components). Please clarify role of 'platform' to get appropriate answer.
Chief Information Technology Officer in IT Services12 days ago
for us initial setup is heavy and can lead to resistance if devs feel their autonomy is being replaced by rigid controls.
We are gradually moving towards platform engineering, although it is not through a transformation initiative so things are more gradual. That being said, it's been ongoing for several years, so I can provide some gotcha's and benefits...
A main benefit we see, if applied successfully, is that the platform team is less and less involved in the business IT initiatives, freeing up time and resources to truly focus on the quality and efficiency of the platform itself. In the past, we noticed that underlying technologies got insufficient attention because the teams were constantly involved in supporting business IT teams. While from a value perspective, the business initiatives are of course important, the lack of a stable and foundational technology base to rely on causes issues later down the road, impacting also customers and perception.
Second, the expertise that platform teams need are strongly focusing on non-functionals, operational automation, infrastructure security, etc. This is a different profile than from a development perspective. For areas where we have a platform-less DevOps approach, we find that it is very challenging to get that expertise within the DevOps team, whereas for platforms teams (and PlatformOps) thus is much more prevalent.
By focusing on platforms, we can also streamline our architecture and application landscape, and can provide a more simplified view on what IT entails. This not only benefits our internal customers (better knowledge of the environment in which they'll be active) but also regulators, auditors, management, etc.
The downside we see is that it is very slow in adoption, mainly because this isn't tackled through a transform initiative but rather by nudging projects left and right. But that also means that not all aspects of platform engineering are continuously approached: people (expertise, training, coaching, culture), process (decoupling from vertical stack to platform-application, IT process alignment, standardization), technology (standardization of technologies, finding the best technology, abstracting application reliance on specific technologies and move to more interoperable hosting).
Furthermore, not all applications are suitable to be considered in a platform-application model, but large organizations have a hard time looking into bimodal setups. You start explaining benefits of platforms, and the first question you get is how this applies to a legacy and monolithic application with its own technology stack completely independent of the 1799 other applications your environment supports.