Edgelixir: Distributed Graph Processing in Elixir
summary
Edgelixir implements a Pregel-based model for distributed graph algorithms using Elixir and OTP, with graph storage via digraph backed by ETS, global name registration and a GenServer barrier to coordinate supersteps. It partitions by vertex ID (edge-cut), spawns a process per vertex for compute, and prints results to stdout. The project is early and practical gaps remain: no benchmarks and no checkpointing for fault tolerance, but the core message-passing and compute loop run as described.
AI-written summary · about
chapters (6)
- 0:00Start
- 0:06Platinum
- 2:11Talk Outline
- 12:16The Toolbox
- 23:17Superstep Supervisor: Message passing
- 24:22Vertex Processes
related talks
39mLearn Elixir by Building a Real Time CLI Chat App (Part 3)
45mLearn Elixir by Building a Real Time CLI Chat App (Part 1)
24mLearn Elixir by Building a Real Time CLI Chat App (Part 2)
15mEfficient GPU-Centric Evolving Graph Processing at Scale
25mApplied Machine Learning in Elixir
25mArtificial intelligence with Elixir the Rete (...)
20mScaling Edge AI with Elixir and Nerves
39mIntroducing Elixir: Your entire web stack