SDDC storage capacity and data redundancy scale with the number of nodes in the SDDC.
Single node SDDCs provide no data redundancy. SDDCs with more nodes support data redundancy through various RAID configurations. All SDDC storage provides data deduplication and compression. Data redundancy in SDDCs that have more than one host is expressed as a number of failures to tolerate (FTT), where a failure can be the loss of a single host in a cluster or a single storage device in an array.
Typical Storage Allocation and Consumption Scenarios
All RAID configurations consume data to support redundancy, as shown in this table. The Usable Storage column shows how much data you could store on a given host under a specific vSAN policy, and accounts for recommended free space by subtracting 30% from that total.
|Nodes||FTT and RAID Config||Usable Storage (TB)||Projected Usable Capacity After Dedup/Compression (TB)|
|3||1 failure, RAID-1 (Mirroring)||3.7||5|
|4||1 failure, RAID-5 (Erasure Coding)||5.6||8|
|5||2 failures, RAID-1 (Mirroring)||2.5||3.6|
|6||2 failures, RAID-6 (Erasure Coding)||4.9||7|
RAID overhead can also be expressed in terms of the storage consumed by a VM. This table shows how various RAID and FTT configurations affect the average amount of storage consumed by a 100 GB VM for a given storage allocation.
|Nodes||FTT and RAID Config||Actual GB Allocated||Average GB Consumed|
|3||1 failure, RAID-1 (Mirroring)||200||138|
|4||1 failure, RAID-5 (Erasure Coding)||133||92|
|5||2 failures, RAID-1 (Mirroring)||300||207|
|6||2 failures, RAID-6 (Erasure Coding)||150||103|