CSM microservices were generally designed and built four to five years ago. The choices made about which libraries to include and what paradigms to embrace were in the context of tradeoffs based on the information at the time. As part of a fully integrated solution, little emphasis was put on the independent functioning of each microservice and release cycle of each repository. In a more modular future for ochami, we must re-evaluate some of these choices with current context. Some decisions will hold because they remain the best possible solution. Other choices will hold because they remain the optimal solution with current information, even if "better" options are available. Developer time is not free and good enough is better than perfect in most cases. There will also be a set of choices that we evaluate as no longer appropriate for the project. Simply put, as CSM has been developing, the world of open source software has been innovating.
As we work on the microservices to make them more independent, we must review the go dependencies, the software boundaries, and even the landscape of alternatives to assess the highest leverage action we can take to make each microservice most valuable to the operators and users of the HPC systems.
CSM microservices were generally designed and built four to five years ago. The choices made about which libraries to include and what paradigms to embrace were in the context of tradeoffs based on the information at the time. As part of a fully integrated solution, little emphasis was put on the independent functioning of each microservice and release cycle of each repository. In a more modular future for ochami, we must re-evaluate some of these choices with current context. Some decisions will hold because they remain the best possible solution. Other choices will hold because they remain the optimal solution with current information, even if "better" options are available. Developer time is not free and good enough is better than perfect in most cases. There will also be a set of choices that we evaluate as no longer appropriate for the project. Simply put, as CSM has been developing, the world of open source software has been innovating.
As we work on the microservices to make them more independent, we must review the go dependencies, the software boundaries, and even the landscape of alternatives to assess the highest leverage action we can take to make each microservice most valuable to the operators and users of the HPC systems.