As processors exchange messages, they gain more information about the overall system state. Mathematically, this acts as a of the input complex.
Because an asynchronous execution complex is continuously connected, and a continuous map cannot map a connected space onto a disconnected space without tearing it, . You cannot map a space without holes onto a space with holes without breaking the rules of the protocol. Key Frameworks and Variations
: Proving a task is impossible requires showing that a certain topological map does not exist (e.g., trying to map a sphere onto a circle without tearing it). Key Textbooks and Foundational Literature
The central breakthrough of this field is the ability to transform (which unfold over time with unpredictable delays) into static combinatorial structures . distributed computing through combinatorial topology pdf
: Represent the local state of a single process (what it knows).
The fundamental building block is a (plural: simplices). A 0-simplex is a single point (vertex). A 1-simplex is a line segment connecting two points. A 2-simplex is a solid triangle.
Because the protocol complex is (specifically, it lacks operational holes), and the output complex for consensus is disconnected (divided into distinct "all-decide-0" and "all-decide-1" regions), no continuous mapping can stretch the protocol complex onto the output complex without tearing it. This topological mismatch provides a geometric proof of consensus impossibility. 4. The Wait-Free Solvability Theorem As processors exchange messages, they gain more information
. It provides a unified framework to replace scattered conference papers with a standard terminology for analyzing algorithms in multicore processors, wireless networks, and internet protocols. Amazon.com Core Concepts and Methodology
The topological approach offers a powerful solution by condensing all possible executions into a single, static simplicial complex , a kind of multi-dimensional shape built from vertices, edges, triangles, and their higher-dimensional analogs. This structure is called a .
Herlihy, M., Kozlov, D., & Rajsbaum, S. (2013). Distributed Computing Through Combinatorial Topology . Morgan Kaufmann. You cannot map a space without holes onto
Because each vertex belongs to a specific processor, the complexes are . Every vertex is assigned a color representing a unique processor ID. No two vertices in the same simplex can share the same color, ensuring that a single global state never assigns two different local states to the same processor. Combinatorial Subdivision
In combinatorial topology, the fundamental unit is a .
: Rounds of communication "subdivide" the input complex into smaller pieces. If the resulting complex remains "well-connected," certain tasks (like Consensus ) may be impossible to solve because processes cannot "break" the connectivity to reach a single decision.
One of the key ideas in the book is that of the . Instead of enumerating every possible execution path, combinatorial topology allows us to represent the entire set of executions of a distributed algorithm as a single, static mathematical object: the protocol complex. The structure of this object—its holes, connectivity, and higher-dimensional properties—directly reflects the solvability of a computational problem.