How Factory Approach to cloud adoption yield best outcome?

In advent of AI, every organization is setting up goal for next 2-3 years to move their compute and loads in cloud. Even multi cloud is something everyone is looking for. Having a multi cloud approach help in diversification of risk. In all of this the number of projects of cloud adoption would be on the rise.

Opportunities to migrate existing infrastructure brings in new challenges. Based on my experience in Cloud Adoption projects the Factory Approach yield best outcome. The term "Factory Model" in the context of software development is often associated with a software development methodology or approach that emphasizes efficiency, standardization, and systematic processes, similar to how a factory operates in manufacturing.

The existing on-premise infrastructure is vast and spread across boundaries. The biggest challenge is there is no pattern in existing on-premise infrastructure. Originally developed and maintained by multiple vendors, there are many hands involved. We have seen a situation when client throw a git repository to us and say this is all we have. In cloud adoption projects situation is such worse that we don't even realize, and project turn from migration to redevelopment.

I would like to share one of my recent experiences. We were planning to migrate around 250 jobs into AWS Glue. Although the team was familiar with AWS Cloud the tool was new. Only few folks were aware of the AWS Glue. The most important thing we were racing against time. There was supposed to be no learning curve. In this situation we decided to adopt Factory Model Approach.

Standardization is the key in Factory approach. We did a factory setup of migration. First, we did initial requirement analysis. In requirement analysis we put the most experienced people. The target was to find all possible requirements for the existing jobs and document this. Categorize the overall work in different buckets. Prioritize the buckets based on the business requirements and complexity. Lined up this bucket of work with a mix of complexity and easy work. Whole plan was set in phases. Initial phase were a combination of unknown bucket of work and easier one to showcase and gain confidence form stakeholder. Subsequent buckets were lined up with flexibility of change of scope. This finalized the blueprint of the factory.

Once this ready it's time to set the processes. We identified pattern if any. In our experience we always get a pattern in work. Again, we put a small team of experts who identified pattern and put the job as a framework. We found that almost 50 jobs are similar in nature. We created a reusable code which team can easily pick. Team need not know AWS glue. They just knew the process and way to test. Problems were common and were resolved across the board.

The factory approach reduces the time to deliver. We promised 30 jobs in one week and delivered almost thrice of that. The most important aspect we gained is the scalability. We could easily add or reduce the number of people aligned to this process. Giving more control on the entire operation.

In second and third iteration we even automated some of the processes. This reduced the time and total cost of delivery dropped in subsequent phases improving the overall margin. Increased customer confidence is another intangible benefit we received.

Factory Approach is cloud adoption is amazing approach. I have adopted this almost four times till now. Every time it best results.

Thanks for Reading!! Happy Learning!!