Designing the Data Structure › Choose to Distribute or Replicate Data › Challenges of Distributed and Replicated Data › Integrity of Replicated Data

Integrity of Replicated Data

With replicated data, the data is replicated as identical copies on the clients, or possibly on a number of distributed servers.

Integrity considerations for replicated data integrity:

• How will the data be updated?
• What mechanisms will be used to ensure that each client has access to the same data at all times?
• Where is the master copy of the data?
• Is the data needed immediately, or is a snapshot sufficient?
• Will the whole table be replicated or just a subset?

Note: Ideally, replication should be used only to give read-only processes better performance.

Updating replicated data should be avoided as it is a very high risk process. If replicated data is updated, the following minimum procedure is required:

1. Record the source of the data.
2. Update the source.
3. Apply the updates to the replicated data.