We are glad to announce the release of GraphScope 0.6. This major release integrates a new graph interactive engine GAIA, which supports efficient parallel execution and bounded-memory execution for Gremlin queries. More technical details of GAIA can refer to our published tech blog. Note that currently the integration of GAIA with GraphScope is experimental, and is not recommended for production use yet! In addition, this release improves the experience of local deployment on MacOS, Ubuntu and CentOS, and adds more graph analytics algorithms.
We highlight the following improvements included in this release:
- [Experimental] Integrate GAIA, a graph interactive query engine, into GraphScope. Currently, it has supported the following features/functions:
- Dynamic memory management for arbitrary graph traversal with ensuring bounded use of memory;
- Automatic and adaptive strategy for optimizing Gremlin traversal, such as hybrid DFS/BFS traversal to balance parallelism and memory usage;
- Early-stop optimization for Gremlin (limit, nested conditional, etc.) to minimize wasted computation;
- Improvement of performance and scalability (a new LDBC Social Network Benchmark will be released around year end);
- Support both Vineyard and the new persistent graph store.
Lazy evaluation support for graph interactive engine and graph learning engine.
A script supporting local deployment on MacOS, Ubuntu and CentOS.
- Add more graph analytics algorithms as built-in applications.
For more detailed improvements that have been made in this release, please refer to the complete changelog.