Growing Order Data – Is this really an issue?
The IBM Order Management (OM) database grows as sales increase and as new business units and channels are on-boarded to utilize a common business process. Usually, an upsurge in transactional data volume is observed over time and this growing data becomes a problem for customers. The sales order data can grow significantly large for relational databases to provide an optimal response. A large data volume not only slows down the “read and search mechanisms” but also contributes to degraded performance of other transactions like “inserts and updates”.
In order to continue taking orders at a faster rate and address the degraded performance, “read” operations can be moved from the main transactional database to an efficient and more agile datastore. The active transactional database can then be trimmed down to house relatively small transactional data. For example, retain transactions that are still in fulfillment process or transactions that are based on certain pre-defined criterion, which is justifiable with business use cases. This approach not only improves the performance of the “reads” but also improves the “inserts and updates” to the transactional database with reduction in the data volume.
Data that is stored in an external datastore is still accessible as readily as “online data” and benefits over purging and storing data on tape. As an option, this external datastore can optionally be subdivided. For example, by type of data or by time (sales orders older than ‘X’ years).
A similar approach can be used for other data sources like main frames to provide an enterprise-wide order search service.
The following diagram shows a high-level view of putting this all together to manage growing data in IBM Cloud.