Even having all required information in a single data repository again was good progress the next challenges were to put all data into relationship, build the desired reports and deliver them to the users with no or minimal effort.
Aggregate, concentrate and summarize large amount of acquired data
In addition it was necessary to keep an eye on potential performance and resource issues caused by the high amount of data received from the shop floor. The fact that every machine delivers 6 records per minute creates an amount of almost half a million records for 50 machines per day! This definitely is ‘Big Data’ which, without any form of aggregation would quickly have led to serious problems.
Again Oracle’s Manufacturing Operations Center (MOC) provided the right tools to get this job done. In configurable intervals MOC reads all data from the so called ‘raw data’ and ‘staging’ tables and aggregates it into several layers of summary tables each with another purpose for example machine-states, output quantities, downtime or scrap reasons, operational or quality data. After processing the raw data MOC makes room for new data by cleaning up the initial table. To further increase performance, data in the summary tables is made available to the reporting components via so called ‘materialized views’.
Reading and validating of raw data followed by building summary tables and materialized views
Automatic cleanup of raw data after processing
Reprocessing of incorrect input data
Reduced hardware costs by compression of data
Ensured performance by aggregating raw data into summary tables and materialized views
Enabled reprocessing of invalid or forgotten input data