Ultra-large-scale graph data storage
A single graph can store hundreds of terabytes of data. A Graph Compute cluster can store tens of billions of vertices and hundreds of billions of edges.
Various indexes
Graph Compute supports multiple types of indexes including key-value (KV) indexes, PKey-SKey-Value (KKV) indexes, built-in indexes for text retrieval, and vector indexes.
High-performance operators
Graph Compute is compatible with Gremlin operators and seamlessly integrated with the graph computing ecosystem. Graph Compute provides multiple optimization solutions, such as computing on storage nodes and refactoring for specific operators. The query performance of Graph Compute is more than five times better compared with the best solution in the industry.
Multiple supported data sources
Graph Compute allows you to import source data from MaxCompute, Log Service, and Apsara File Storage for HDFS with a few clicks.
Graph Compute allows you to migrate data at a fast speed from Neo4j and Realtime Compute for Apache Flink by using a dedicated tool.
Flexible data management
The offline system is asynchronously built and supports full data switching at hourly intervals.
Graph Compute provides maintenance for multiple data versions and allows you to roll back data within minutes.
The fine-graded time to live (TTL) management allows you to manage the lifecycle of a single data record.