一种支持复杂关系查询的图存储系统

A Graph Data Store with Complex Query Support

大数据时代,数据间的高相关性为大数据带来了重要价值,"图"这一抽象数据结构能够很好地表示数据间的关系.复杂图数据具有数据规模大、建模需求多、关系查询复杂等特点,为图存储系统的构建带来挑战.提出了一种支持时间演化和数据溯源的图数据模型,形式化定义了数据结构、数据操作和数据完整性约束,为大规模图数据存储系统的实现奠定了理论基础;实现了一个支持复杂数据查询的图存储系统,为分析挖掘图数据的巨大价值提供了底层基础设施技术支撑.通过实验对系统的存储、入库、查询等能力进行了效果评估,结果表明所提方法有效可行.

The highly connected big data bring us big opportunities to create value.The abstract data structure'graph'is well-suited to model the associations between data elements and represents an important form of big data.Existing complex graphs share the characteristics such as huge data volume,complicated modeling requirements and complex queries.Therefore,we propose a graph data model with time dimension and data lineage support.We give both formal description and intuitive examples for the model and define the semantics of graph queries,which provides a theoretical foundation for our system.We build a distributed graph data store with complex queries support,which provides technical support for analyzing the great value of graph data.Experiments on large graph datasets show its good read/write performance.