Distributed Keyword Search over Graph Databases using IBM Analytics Platform

世界上许多高质量的企业和社交数据以半结构化和结构化数据的形式存储，其中包括企业的RDBMS、知识图谱和社交网络。所有这些数据集合要么已经定义为图形数据库，要么可以重新建模为图形。在过去的十年里，我们见证了存储和操作结构化和半结构化数据的进步，但我们并没有看到在搜索方面的太大进步。正如Surajit Chaudhuri(微软研究院的一位杰出科学家)在2015年IEEEData Engineering Conference上的主题演讲中所说的那样，基于结构化数据库的搜索已经落后于基于非结构化数据的搜索。当科学家和商业用户从他们的不同数据集中寻找令人兴奋的、可操作的发现时，提供有效搜索的需求是深远的。如果我们不能提供一个强大的搜索系统，许多大数据的梦想就不会实现。与一组完全不同的、非结构化的文档相比，图形数据库中的结构的存在为数据探索提供了强有力的手段。传统上，要访问结构化数据库，用户必须学习结构化查询语言，如SQL或SPARQL。他们还需要学习他们感兴趣的数据库的模式。然而，非技术用户(即任何不懂查询语言或不熟悉给定模式的人)实际上被锁在了门外。IBM的Analytics平台目前不支持非技术用户探索图形数据库。这个项目将为这个问题提供一个解决方案。更具体地说，我们致力于利用IBM Analytics平台设计一个分布式、并行的关键字搜索系统，使非技术用户能够探索图形数据库。

Much of the world's high-quality enterprise and social data are stored as semi-structured and structured data.This includes enterprises' RDBMSs, knowledge graphs, and social networks. All these data collections eitherare defined already as graph databases or can be re-modeled as graphs. Over the past decade, we havewitnessed advances in storing and manipulating structured and semi-structured data, but we have not seenmuch progress in search over them.As Surajit Chaudhuri (a distinguished scientist at Microsoft Research) addressed in his keynote talk at IEEEData Engineering Conference in 2015, search over structured databases has fallen behind search overunstructured data. While scientists and business users look for exciting, actionable discoveries from theirheterogeneous datasets, the need to provide effective search is profound. If we cannot deliver a powerful searchsystem, much of the big data dream will not be achieved.The existence of structure in graph databases offers potent means for data exploration compared with a set ofdisparate, unstructured documents. Traditionally, to access structured databases, users have to learn structuredquery languages, such as SQL or SPARQL. They also need to learn the schemas of the databases in which theyhave an interest. However, a non-technical user (i.e., anyone who does not know query languages or is notfamiliar with the given schema) is effectively locked out.Exploring graph databases for non-technical users is currently not supported in IBM's Analytics Platform. Thisproject will provide a solution for this problem. More specifically, we focus on designing a distributed andparallel keyword search system using IBM Analytics Platform to empower non-technical users to explore graphdatabases.