How is data stored redundantly in a zone-redundant storage account?

Question

Hi all,

I am curious about how data is redundantly stored in a zone-redundant storage (ZRS) account on Azure. I came across the SOSP paper, which described the old architecture that did not support ZRS. I understand that in the old architecture, data was replicated and stored in three copies in a stamp. Does ZRS implement synchronous replication between 3 stamps in each availability zone, or a stamp that lives across 3 availability zones? Does Azure keep additional copies of the data in each zone?

Capacity-wise, would LRS and ZRS general-purpose v2 accounts share stamps, thereby sharing the available capacity in each zone?

Answer 1

@ML-1722 Welcome to Microsoft Q&A Forum, Thank you for posting your query here!

There are multiple replicas that are synchronously replicated across the AZs. LRS and ZRS do not share stamps. 

If you still have any question, please let me know I would like to work offline on this issue.

---

Please do not forget to "Accept the answer” and “up-vote” wherever the information provided helps you, this can be beneficial to other community members.

Answer 2

As per https://learn.microsoft.com/en-us/azure/storage/common/storage-redundancy#zone-redundant-storage , ZRS implements synchronous replication across 3 stamps, each in a different availability zone

Regarding capacity sharing, tbh, I don't see how that's relevant. Ultimately, that's abstracted away by the platform - ultimately, as a customer, you rely on the corresponding SLAs/performance targets. In short, I don't believe this information is available publicly.

---

If the above response helps answer your question, remember to "Accept Answer" so that others in the community facing similar issues can easily find the solution. Your contribution is highly appreciated.

hth

Marcin