A large-scale graph computing system, basic on disk/memory & integrate with graph database HugeGraph

Clone this repo:
  1. 1800517 feature: optimize data receiving (#181) by Cong Zhao · 5 months ago master
  2. bcaef85 feature: support partition concurrent compute (#177) by ShouJing · 5 months ago
  3. 0d2e6df rebuild ci to compatible apache (#184) by Cong Zhao · 5 months ago
  4. 50b6147 chore: support codeQL check (#183) by imbajin · 7 months ago
  5. f5c7c48 chore: use cla assistant to support robot pr (#182) by imbajin · 7 months ago


License Build Status codecov Docker Pulls

hugegraph-computer is a distributed graph processing system for hugegraph. It is an implementaion of Pregel. It runs on Kubernetes or YARN framework.


  • Based on BSP(Bulk Synchronous Parallel) model, every iteration is a superstep.
  • Auto memory management. The framework will spilt some data to disk, the framework will never OOM(Out of Memory).
  • The the part of edges or the messages of super node can be in memory, so you will never loss it.
  • You can output the result to HDFS or HugeGraph, or any other system.
  • Easy to develop a new algotirhm. You need to focus on a vertex only, not to worry about messages transfering and memory.

Learn More

The project homepage contains more information about hugegraph-computer.


hugegraph-computer is licensed under Apache 2.0 License.