| In the increasingly popular trend of big data, the need for massivedata processing is becoming bigger than ever. Applications rely ontraditional single instance central relational database may encounterscalability and performance issues. Distributed cloud database with highscalability and performance can solve these problems princely well.Original legacy applications can achieve these advantages throughdatabase migration from central database system to distributed clouddatabase system. However, this migration process is relatively complex.Data migration is only part of this process and is relatively easy. What’s alot more complex is the migration of large number of user defined storedprocedures.Since conventional stored procedure is defined and executed within asingle database, it is not suitable for distributed database under multi-nodescluster. Expected results cannot be achieved by simply storing andexecuting it. This study provided a mechanism for migration of storedprocedures from conventional single instance database to a distributedrelational cloud database. This cloud database is based on sharding pattern.Under this mechanism, stored procedure source code was parsed andtranslated to another form which could be executed by the engine of thetarget cloud database. Execution of database accessing statements wasdelegated to the engine, whose results participated in the subsequentexecution of other statements.This study and its implementation mainly aimed at the popular OraclePL/SQL subprograms. Migration target was a distributed cloud databasefrom Intple Technoloby, called DBOne. The implementation of this migration system enabled the correct creation, store and execution ofPL/SQL subprograms within DBOne, facilitating the database migrationprocess. Beyond the specific migration source and target, the mechanismoffered by this study can also be used for reference for other databasesystems. |
PDF Full Text Request