Replication of storage represents how the blobs are replicated to ensure the safety of their contents in case of hardware failure. When we create a blob, one of the options we have to select is the replication type:

• LRS (short for Local Redundant Storage): This storage replicates each blob three times in the same data center. It is designed to give us 99.999999999% (eleven 9s) in terms of durability of our objects. This is the least expensive option but offers less durability than other options.
• ZRS (short for Zone Redundant Storage): This storage replicates each blob in different zones. It is designed to give us 99.9999999999% (twelve 9s) in terms of durability of our objects. It's used to avoid downtime for applications that use them.
• GRS (short for Geo-Redundant Storage): This storage replicates each blob across Azure regions. Once a blob is written in the primary region, it is replicated asynchronously in a region that might be hundreds of kilometers away. This storage manages to give us 9.99999999999999% (sixteen 9s) in terms of durability. It allows making our data available even if a disaster happens in the primary region.
• RA-GRS (short for Read-Access Geo-Redundant Storage): Similar to GRS, the difference is that the secondary region is made available in read-only mode. This makes applications that use it much more scalable in term of read access performance.

All these options are available when we create a blob storage account. The more durability we have, the more it will cost. The following URL gives an overview of the costs associated with these storage options: https://azure.microsoft.com/en-us/pricing/details/storage/blobs/.