Last time we started our operation-based CRDTs sub-series, as we moved away from state-based CRDTs. We talked mostly about core requirements and sample implementation of RCB (Reliable Causal Broadcast) protocol, which was necessary to provide guarantees required by Commutative Replicated…
Operation based CRDTs: protocol
Today we'll continue a series about CRDTs, this time however we'll stray from the path of state-based CRDTs and start talking about their operation-based relatives. The major difference that we need to cover, is the center of gravity of this…
Hybrid Logical Clocks
In this blog post we'll going to talk about timestamps. But to start with we need to answer the basic question: what are we using timestamps for? Two most common cases are: We want to timestamp our records so later…
Make your cluster SWIM
In this blog post we'll cover how systems form clusters, what clusters actually are and what are their responsibilities. We'll also present different protocols responsible to serve the needs of the clusters with a various tradeoffs associated with them. To…
Dealing with complex dependency injection in F#
Today, we're going to cover different ways of encapsulating capabilities and supplying them between functions using functional programming techniques which can be realized in F#. Managing code dependencies in object oriented languages in 2020 is pretty much one sided problem:…
State-based CRDTs: Maps
In this blog post, we'll cover the idea of CRDT maps, and how we could create them and utilize them in common scenarios. A prerequisite for this talk is some general knowledge of CRDTs, especially observed-remove sets, which were already…
Building custom fibers library in F#
Over the course of last few months on this blog post, I've been sharing about internals and how-to of different concurrency patters. We discussed how to implement our own actors and specific affinity-based thread pool. Today we'll focus of the…
Thread safety with affine thread pools
In the previous blog post I've written about building a thread-safe state access by using actor programming model (as opposed to common practices requiring synchronization barriers with locks) without compromising performance coming with running our programs on multi-CPU machines. As…