A large educational Institution recently approached us with a problem. Their Student Information and Financial databases were in Oracle, and a data warehouse containing extracts from both of these used for decision making, including opening more sections of a class or cancelling a class, was in SQLServer. Extract-Transform-Load process extract data from the Oracle databases, create “staging tables” in SQLServer, and then transform and load processes make it available to the decision makers.
The current extract process was taking over 8 hours, which was cutting into their “batch window” at night, and leaving decision makers with data at least 24 hours old. One of their staff attended a WebEx on SharePlex and wanted to understand how Quest might be able to assist in solving this problem
SharePlex to the Rescue
The customer set up three SharePlex replication streams, one from each of their Oracle sources, to their SQLServer target.
Taking advantage of SharePlex’s built-in partitioning, they were able to select only the rows and columns they wanted, cutting down dramatically on the amount of data transferred between systems. Since SharePlex captures data from the Oracle redo log “as it happens”, they were able to present current data to the transformation and load processes on the SQLServer systems. Those processes then prepare the data for the decision support system.
By using SharePlex to eliminate the extract piece of their Extract-Transform-Load process, our customer cut the lag in the availability of data to decision makers from over 8 hours to less than 2.