All production briteVAULT clusters are built using a highly-available architecture.  All compute nodes in the cluster are deployed with N+1 capacity, meaning a cluster node can fail and the remaining nodes have enough RAM and compute to continue to run the production workloads. 

​

All production storage is also highly-available and fault-tolerant all the way up to the chassis.  This means that each production cluster has multiple independent storage chassis, and the storage fabric can suffer a drive, controller, or entire chassis failure and production workloads continue to operate.  Storage fabric communication is multi-homed on active/active links, so if one storage path fails another is available.