Heterogeneous-operator management method in big-data analysis and development platform

TL;DR: In this article, a heterogeneous-operator management method for big-data analysis and development platform is presented, where operators are constructed as nodes, a node running instance is generated by the nodes to provide an interface and corresponding bottom-layer implementation, good support is provided for an analysis platform, a function of archiving arrangement on the different-type operators is implemented, development time is reduced, management interfacespecifications of the operators are unified and maintenance and updating on an operator library are facilitated.

Abstract: The invention provides a heterogeneous-operator management method in a big-data analysis and development platform. The method includes: S1, encoding various algorithms, which are used for data analysis, to generate operators, and saving information of the operators into a database; S2, using an operator new-adding interface to save a customized operator into the database according to information which is of the customized operator and is imported by a user; and S3, using an operator query interface to query an operator in the database according to filtering parameters in a query request when the query request is received. According to the method, the operators are constructed as nodes, a node running instance is generated by the nodes to provide an interface and corresponding bottom-layerimplementation, good support is provided for an analysis platform, a function of archiving arrangement on the different-type operators is implemented, development time is reduced, management interfacespecifications of the operators are unified, and maintenance and updating on an operator library are facilitated. The method is easy to expand, and is high in adaptability.