More than 12 years of experience, currently usually as architect/(lead) developer – but my roles vary from analysis to deployment. Interested mainly in JVM (Scala in particular) My main fields of interest are functional programming, stream processing systems (Akka, Kafka, Flink) and integration.
I also like to give talks at conferences – Confitura, HeapCon, Devoxx just to name a few. Currently I’m working on TouK Nussknacker project which enables analysts to create streaming jobs with friendly UI.
When I was young and took my CS classes I learned about ACID and it seemed sound and reasonable. Then I went to work and learned that it could mean different things (especially isolation…) for different database vendors…
Around 2009 I’ve heard about CAP theorem - that was also sound and quite reasonable in the brave, new distributed world. And then one after another - NoSQL vendors started to claim they’ve beaten CAP theorem (* as they understand it of course…)
And recently it got even more complex - stream processing engine have their notion of transactions - atomicity, isolation and so on.
In this talk I will try to make something out of this mess - we’ll look at some of definitions and try to figure out which use case require particular kind of isolation or transactionality. Of course some of it will be based on my own experience - including quite a lot of blood, sweat and tears ;)
Text box item sample content