User Experience

The Native Cloud Object Storage support provides a simple path for adoption through its direct integration within the existing storage hierarchy of Db2, allowing existing applications to leverage the low-cost object storage for Db2 tables without changes.

To create a table stored in object storage, users will follow the existing process of creating a table in Db2. Tables are created within a storage hierarchy, with the object created in a table space, and the table space is associated with a storage group. The main difference introduced by this feature is in the definition of the storage group. It will now be defined as associated with a storage access alias, instead of being associated with a set of storage paths. This designates it as a storage group that is persisted to object storage, or a 'remote' storage group.

Subsequently, any table space associated with that remote storage group will be stored within the storage defined by the storage access alias associated with the storage group. These table spaces are referred to as remote table spaces. Similarly, any database object such as tables, indexes, MQTs, etc., associated with that remote table space will also be stored in the same object storage target and will be able to leverage the performance advantages of the multi-tier cache.

Experience in IBM Db2 Warehouse as a Service in AWS

IBM® Db2® Warehouse as a Service hosted on Amazon Web Services includes the support of Native Object Storage within its managed service and is pre-configured for users to directly exploit this feature. This includes the provisioning of the Object Storage container (bucket) and the configuration of the storage hierarchy, including two remote table spaces - one regular table space OBJSTORESPACE1 and one user temporary table space OBJSTORESPACEUTMP1. These can then be used to create both column-organized tables and column-organized Declared Global Temporary Tables in object storage.