Modernize or rewrite a legacy system? How to decide
Every legacy system reaches a point where the question comes up: modernize what exists, or throw it out and rewrite? The intuitive answer (“rewrite everything”) is usually the most expensive and risky.
Why rewriting from scratch is tempting (and disappointing)
Rewriting looks clean: new code, none of the past’s hacks. But you throw away years of business rules learned the hard way, you stop evolving the product during the rewrite, and you risk recreating the same problems in new clothes.
When modernizing is better
- The system still serves the business but stalls at specific points.
- The business rules are valuable and hard to re-document.
- You cannot stop the operation for months.
When a rewrite is justified
- The base technology has no support and no path forward.
- The cost of maintaining already exceeds the cost of rebuilding.
- The product will change so radically that little is reusable.
The middle ground that almost always wins
In most cases, the best answer is incremental modernization — replacing the system in parts, with the strangler fig pattern, with no blackout. That is how we do maintenance and modernization of critical systems. More on the topic in when to sustain vs. modernize.
Have a legacy that holds the business back?
We modernize critical systems in parts, without stopping the operation.