Introduction to sharding

Sharding separates large databases into smaller, more easily managed parts called database shards that can be spread across multiple database servers.

In IBM® Sterling Selling and Fulfillment Foundation, a colony is a set of database shards and is determined by a configurable set of attributes.

The following diagrams illustrate how several sharding configurations work.

Sharding configuration 1 (sharding by enterprise)

In this diagram, two Transaction shards, Transaction 1 and Transaction 2, share the same Configuration and Statistics shard.

Sharding by enterprise example

In this example, Transaction data for two of the enterprises, E1 and E2, resides in the Transaction 1 database, while Transaction data for Enterprise E3 resides in the Transaction 2 database. Master data is shared by all enterprises. You can create a colony to bound the set of enterprises that share the same Transaction shard.

Sharding configuration 2 (sharding by enterprise and seller)

In this diagram, three Transaction shards (Transaction 1, Transaction 2, and Transaction 3) share the same Configuration and Statistics shard.

Sharding by enterprise and seller example

In this example:

Sharding configuration 3 (sharding by enterprise and seller)

This diagram is very similar to the Sharding Configuration 2 diagram, except that there are two Master databases, Master 1 and Master 2.

Sharding by enterprise and seller, configuration 3

In this example:

Additional sharding considerations

Note the following sharding considerations: