This approach revolutionized how computer scientists reason about concurrency. It moved the field from using graph theory and temporal logic to using .
Last updated: 2025 – This article reflects the current relevance of combinatorial topology in light of new fault-tolerant blockchain protocols. distributed computing through combinatorial topology pdf
In the late 1980s and early 90s, computer scientists Maurice Herlihy, Sergio Rajsbaum, and others asked a bold question: What if we stopped looking at the steps and started looking at the space of all possible outcomes? In the late 1980s and early 90s, computer
Later, Aris explained to a new recruit, pointing at the topology textbook on his desk: "In a perfect world, consensus is easy. But in a distributed system, the set of possible failures creates holes in the logic—holes that topology can see. We don't solve the impossible. We navigate the shape of the possible." We don't solve the impossible
Consider the problem (a generalization of Consensus). In Consensus, all processes must agree on one process's input. In Set Agreement, processes must agree on a set of at most k input values. Proving impossibility for k consensus is trivial; proving impossibility for Set Agreement is not.
Distributed computing through combinatorial topology has a wide range of applications, including:
A represents the collective state of processes.