With the sharply increasing amount of data, studies on bigdata processing based on NoSQL have been actively
done. However, NoSQL cannot satisfy the ACID properties of database transactions. Therefore, bigdata processing
based on RDBMS has been spotlighted. CUBRID Shard stores data in the distributed CUBRID servers by dividing
the database. However, CUBRID Shard cannot process a query when data of a user is distributed on the multiple
CUBRID servers. Therefore, in this paper we propose a CUBRID based middleware which supports distributed
parallel query processing. Our proposed middleware can support users who are familiar with SQL to conveniently
process the bigdata by using SQL statements. In addition, the middleware can support various aggregation operators.
Through the performance evaluations, we show that our proposed scheme shows better performance than the
existing work in terms of query processing time.
As a future work, we plan to expand our middleware to support various types of join with reasonable efficiency