01
Assess what is really there
Dependency and data-flow map, dead code, the parts under test, and the parts only one person understands. Usually the first accurate picture the team has had in years.
Phase 1 · assessment, plan, and one module live
The system nobody wants to touch, made touchable.
A Rails 4 app, a PHP monolith, or a .NET service that still runs the business and that everyone routes around. Phase 1 ends with a plan you believe and one module already migrated and serving live traffic.
We do not propose a rewrite. Rewrites of systems whose behaviour nobody fully knows are how companies spend two years and ship nothing.
$20,000
Senior-weighted. Junior time is not useful on unfamiliar legacy code, so we don't bill it.
- Duration
- Phase 1, ~10 weeks
- Team
- Senior-weighted, 2–3
- You provide
- Repo access + a domain expert
- Next slot
- October
What Phase 1 covers
02
Pick the seam
Which module leaves first, and why that one. We pick for a clean boundary and real business value, not for whichever code is ugliest.
03
Strangle it
The new module runs behind a facade beside the old one. Traffic shifts gradually and rolls back instantly. The legacy system keeps running the business the whole time.
04
Hand over the pattern
Phase 2 onward is repetition of a route your own team has now watched work once. Many clients continue without us, which is the point.
What you have at the end of Phase 1
- One module in production. Migrated, serving real traffic, with the rollback path tested rather than described.
- A costed migration plan for everything remaining, in engineer-days, ordered by risk and by what the business actually needs next.
- Characterization tests around the seams we touched — the safety net that makes every later phase cheaper.
- An honest verdict on what should never be migrated at all. Some of it should simply be retired, and someone has to say so.
Ruby on RailsPHP.NETDjangoPostgreSQL · MySQLDockerAWSStrangler fig
Worth knowing
“The first module going live behind a facade is what changes the conversation. After that it is not a bet any more, it is a schedule.”
Assessment · seam · one module live
1
module live before you commit further
0
big-bang cutovers
Questions we get every time
Why not just rewrite it?
Because nobody can tell you everything the current system does, and a rewrite has to match all of it before it can ship. Strangler-fig migration gets value into production in weeks and keeps a rollback the whole way.
Our team wrote it and is still here. Is that a problem?
It is the best case. They know the behaviour we would otherwise have to rediscover, and Phase 1 is designed to hand the pattern to them.
What if the assessment says don't do it?
Then it says that, and you have saved considerably more than $20,000. We have told clients their system was fine and the real problem was elsewhere.
Can you work on a system with no tests?
Yes. That is the normal starting point. Characterization tests around the seam come before any code moves.
Start with Phase 1
Tell us what you’re running. You’ll hear back within one business day, in your timezone, from the engineer who would do the work.
Or WhatsApp +51 914 008 876
Or hello@inspiredsolutionsperu.com