A metadata generation method for massive data in distributed databases

摘要

Being advertisements for source data, metadata is the very portal through which distributed-massive-data applications find data, but metadata generated by traditional methods cannot be automatically updated the time its source data changed. Generally speaking, it is just the data changed that users pay far most attention. So these methods cannot satisfy real-time navigating and instantly providing data requirement from the above applications. To solve the problem, a new method based on database triggers is presented to generate metadata dynamically, also design for data access engine based on metadata is introduced, in which both navigating and monitoring functions of metadata are applied, and it justified that dynamical generation of metadata can satisfied instantly finding and supplying data requirement from real-time distributed-massive-data applications.