In the preceding section we discussed details about the core structure of the Hbase data model in this section we would like to go a step further deep to understand and do things, which will make your data model flexible, robust and scalable. And this is essential to know on an operation level too, as the cluster grows it becomes pivotal to take all consideration.
We should take the following steps to make it happen:
Timestamps
Versions
Sort order
Column metadata
Joins
Timestamps: Every put in HBase is marked with a unique timestamp and this acts as a unique reference for a provided version of a value.
HBase associates timestamp by default mapping it with the respective region server. This default characteristic has a lot of flexibility and you can even specify timestamp value while putting date to the cell of that region server.
In case of multi-datacenter timestamp provides a way to avoid conflicts. In some use cases you can also use it to provide consistency of versions...