Partition tolerance, high availability, coordination avoidance, optimistic UI, offline work lead to conflicts. When timestamp based conflict resolvers do not work, CRDT is a cure.
This talk will demonstrate using a specific usecase and code how you can solve the task of salary and pension payments to millions of clients in a distributed and scalable way.
Developing a tool for detecting potential deadlocks. The talk will cover algorithms, class transformation, performance of the implementation and so on.
Using Spark 2.1 for resilient real-time processing of data coming out of Kafka yearning to be aggregated. During the talk we will fix bugs, edit configs, calculate resource usage and weep.
Examining specifics of client data storage in new Sberbank IT platform.
Architectural challenges, key decisions and our experience designing a corporate system based on in-memory storage.