LinkedIn·Wednesday, 5 August 2026·21d ago
We didn't start by trying to replace application servers. We started by asking a different question. Years ago, many enterprise Java…
Atomikos
1,668 followers
We didn't start by trying to replace application servers.
We started by asking a different question.
Years ago, many enterprise Java applications depended on heavyweight application servers to get reliable transaction management.
Working with customers running platforms like WebSphere and Oracle OC4J made us ask a simple question:
Do you really need the entire application server...
...or do you just need the transaction manager?
That question shaped Atomikos.
Instead of treating transactions as something tied to an application server, we focused on making them embeddable, lightweight, and independent.
Looking back, that design choice turned out to fit much better with today's cloud-native applications than we could have imagined.
Sometimes the biggest innovations begin with a very simple question.
Cross-referenced
Related on the wire
What disappeared when Java moved away from application servers? More than many teams realized. Application servers bundled a lot together:…
A transaction manager does not prevent failures. It determines what should happen when they do. Imagine one business operation needs to do…
Most microservice transaction discussions start with technology. We think there is a better question to ask first: **What happens if this…
Reliable messaging has two sides. Sending reliably means making sure a business update and its outgoing message stay consistent. Receiving…
Most architecture diagrams show the happy path. But reliability is usually decided somewhere in the middle. Consider a simple workflow: -…
Twenty years ago we thought the hard problem was distributed transactions. Today we think it's something else. Over two decades of working…