I'm curious, when you migrate systems that are run in-house to the cloud, which methods do you prefer? For example, lift-and-shift, re-architecting, hybrid, or starting from scratch—what tends to yield the most efficient results? Especially in terms of data security and cost optimization, which steps do you prioritize? What are your general experiences with this?
What approaches are being tried for secure cloud migration?
👁️ 6 views💬 2 replies❤️ 0 likes
2 Replies
I first tried lift-and-shift on a small project using virtual machines, but when moving sensitive data, we realized the cloud provider's security didn’t cover everything we needed. We ended up reworking part of the system to integrate native cloud encryption and cut costs with reserved instances. The most critical part was auditing permissions before the cutover and enabling real-time access logs.
Last year, while working at a mid-sized textile factory, we decided to migrate our aging local servers to the cloud. Our first approach was a *lift-and-shift*—we kept the applications as-is and just moved the servers to AWS. At first, it seemed great because we didn’t have to touch the code, but after three months, the bills skyrocketed. We had over-provisioned resources just to maintain the same performance, and data transfer costs added up fast, eating into our budget.
So, we pivoted to a re-architecting strategy: we Dockerized the core applications and switched to AWS RDS for the database. This time, we saw major improvements in both performance and cost efficiency. We also learned the hard way that security matters—encrypting data and locking down IAM roles were non-negotiable. Funny enough, we never got around to a full "from-scratch redesign," and honestly, with so many legacy dependencies in the old system, it would’ve been nearly impossible. In the end, we went with a hybrid approach: critical apps stayed in the cloud, while local systems ran as backups. Now, we’re in a much better spot with both security and costs under control.