2025-06-13 –, Main Stage - Cinema 10
Designing, verifying, correctly implementing and later improving core distributed protocols like consensus, which are critical for safety and reliability of decentralized systems, is notoriously difficult and error-prone. One of the biggest challenges here is dealing with the inherently concurrent nature of decentralized systems. But it’s too important, and we can’t really afford to get it wrong, so we need to take it seriously and do a much better job here. But how?
In order to make a breakthrough in this area, we would benefit from revisiting the fundamentals and trying to rethink distributed and concurrent programming from the first principles. How should we approach concurrency in distributed algorithms in a natural way and with confidence? How can we avoid excessive synchronization, prevent deadlocks and resource leakage? How can we make distributed and concurrent systems more simple, flexible, and reliable? In this talk, I will share with you some insights into the fundamentals of distributed and concurrent programming that could help us to overcome these difficulties and make a breakthrough for the future of decentralized computing!
As part of my effort to develop Replica_IO, an open-source framework for building practical distributed replication mechanisms, I explored 14 code bases of distributed protocol implementations, 7 frameworks for implementing distributed protocols, as well as a bunch of related concepts, approaches, and techniques. What I realized was that there’s an underlying problem of structuring the inherently concurrent logic of distributed protocols that we don’t really know how to solve in a simple, flexible, and reliable way. Embarrassingly often, we approach this rather awkwardly and unsurprisingly end up with awfully complicated, obscure, and fragile code. So I decided to challenge the status quo and try to rethink the conventional approaches to modeling distributed systems and the way of expressing distributed protocols. And I would like to share the insights that I got so far.
I'm a Research Engineer mostly interested in designing practical, full-featured, composable, and configurable fault-tolerant replication systems, as well as exploring clear and expressive ways to implement them. I'm also enthusiastic about exploring alternative programming models for distributed computing. I think that decentralized Byzantine-fault tolerant mechanisms should dominate in future computing systems, and I'm very excited about working towards that.
I like to think of myself as someone discovering smart solutions to hard problems and doing this for the common good. In that, I believe in open source, open innovation and collaboration.
This is why I decided to found Replica_IO, an open-source framework for building practical distributed replication mechanisms. If you're also excited about this initiative, feel free to contact me! Let's make a breakthrough in designing and implementing distributed replication protocols for the future of decentralized computing 🚀