Stretch clustering with Azure Stack HCI offers great benefits for certain workloads. The following table provides a brief summary for how an organization can benefit from stretch clusters when compared to traditional clusters.
In a traditional cluster, we have the option to create a 2 or 3-way mirror confined to a single site. By contrast, in a stretched cluster each volume on the primary site will require a replica volume on the secondary site and log volumes on both sites. This has profound implications on storage efficiency and performance as we end up with 4x to 6x data resiliency for the data across the two sites.
|Site failure will cause complete outage
|VMs restart on secondary site in case of site outage - Automated failover
|Single Storage Pool
|Two Storage Pools
|2x to 3x data resiliency
|4x to 6x data resiliency
|No performance impact
Performance impact can occur due to inter-site throughput and latency
Additional writes to volumes and their respective log volumes on secondary site can cause overhead
|Complete data loss in case of a calamity
Zero data loss for synchronous replication
Based on RPO for asynchronous