3.2.4 Host and vSphere Cluster Failure Tolerance
vSAN 6.2 introduce a new Storage Policy-Based Management (SPBM) rule, called fault tolerance method, which allows virtualization administrators to choose which method of fault tolerance to employ.
Prior to vSAN 6.2, RAID-1 (mirroring) was used as the failure tolerance method. However, vSAN 6.2 introduced RAID-5/6 (erasure coding) to all-flash configurations. While mirroring techniques excel in workloads where performance is the most important factor, they are expensive in terms of capacity required. RAID-5/6 (erasure coding) data layout can be configured to help provide the same levels of availability, while consuming less capacity than RAID-1 (mirroring). Use of erasure coding reduces capacity consumption by as much as 50 percent compared with mirroring at the same fault tolerance level. This method of fault tolerance does require additional write overhead in comparison to mirroring as a result of data placement and parity.
RAID-5/6 (erasure coding) is configured as a storage policy rule and can be applied to individual virtual disks or an entire virtual machine. Note that the failure tolerance method in the rule set must be set to RAID5/6 (erasure coding).
Figure 7. Storage Policy-Based Management Fault Tolerance Methods
RAID-1 (mirroring) in vSAN employs a 2n+1 host or fault domain algorithm, where n is the number of failures to tolerate. RAID-5/6 (erasure coding) in vSAN employs a 3+1 or 4+2 host or fault domain requirement, depending on 1 or 2 failures to tolerate respectively. RAID-5/6 (erasure coding) does not support 3 failures-to-tolerate. The following table details the host and capacity requirements.
Table 1. vSAN Data Protection Space Consumption
| Tolerated Failures | RAID-1 | RAID-5/6 | Erasure Coding Space Savings vs Mirroring |
| | Minimum Hosts Required | Total Capacity Requirement* | Minimum Hosts Required | Total Capacity Requirement* | |
FTT=0 | 0 | 3 | 1x | n/a | n/a | n/a |
FTT=1 | 1 | 3 | 2x | 4 | 1.33x | 33% less |
FTT=2 | 2 | 5 | 3x | 6 | 1.5x | 50% less |
FTT=3 | 3 | 7 | 4x | n/a | n/a | n/a |
* Without Deduplication/Compression taken into account.
Erasure coding can provide significant capacity savings over mirroring, but it is important to consider that erasure coding incurs additional overhead, which is common among any storage platform for this type of data striping. Because erasure coding is only supported in all-flash vSAN configurations, effects to latency and IOPS are negligible in most use cases due to the inherent performance of flash devices.
