An extension of Two Phase Locking to support parallel updates of data warehouses

Abstract

Even though Two Phase Locking (2PL) has shown a good performance in various applications for on-line transaction processing, it failed to provide analytical transactions with fresh data in data warehousing systems. One of the main reasons for such poor freshness is due to the lack of parallel updates of data warehouses. If several update transactions in 2PL are allowed to execute concurrently without an adequate concurrency control, data inconsistency between data warehouses and operational databases inevitably arises due to incorrect propagation of the updates on the operational databases into the data warehouses. To prevent such an incorrect propagation of update transactions, we extended 2PL so as to process update transactions in an order of their creation if they conflict with each other. In addition, to speed up analytical transactions, locking rules of 2PL are slightly modified in a way of minimizing conflicts among transactions. To investigate the applicable areas of our algorithm, its performance is evaluated by means of simulation approach. Our experimental results show that the extended 2PL enables OLAP transactions to read fresher data than the conventional 2PL.