I85PREP
⌘ K
Dell PowerMax Operate study materials
Dell PowerMax Operate study resources

Dell PowerMax Operate: PowerMax Architecture and Hardware Components

Dell PowerMax Operate · Work through your exam, one topic at a time.

All guide sections

PowerMax Architecture and Hardware Components

The architecture domain checks that you can describe how a PowerMax array is physically built and how the logical objects you use every day map onto that hardware: engines, directors, ports, drives, Storage Resource Pools (SRPs), Service Levels (SLOs) and workload types. Questions often test redundancy and the thin, pool-based capacity model.

Hardware layout

  • PowerMax is end-to-end NVMe: NVMe flash drives in 25-slot disk array enclosures (DAEs), NVMe back-end and NVMe front-end protocols. There is no SAS or SATA tier.
  • The building block is the engine: two directors (A and B), each with redundant nodes, its own NVRAM for write caching, front-end ports and back-end connectivity. An engine plus its DAEs is a brick.
  • The 2500 model scales to two engines; the 8500 scales to eight. Losing a director node, a PSU, a fan or a single drive must not stop I/O.
  • Vault drives and NVRAM hold write cache during a power failure so that acknowledged writes are not lost.

Logical layer

  • An SRP pools the capacity of all drives in the array and is the unit you provision from. An SRP is not a RAID group, and you do not bind volumes to physical drives.
  • SLO (Diamond, Platinum, Gold, Silver, Bronze, Optimized) and workload type (OLTP, DSS and the _REP variants for replicated devices) control how the array services a storage group's I/O. SLO is a property of the storage group, not of an individual volume.
  • Volumes are thin: capacity is consumed as hosts write. Compression is standard; deduplication is model- and release-dependent, so verify before quoting a data reduction ratio.

Worked example

A database will be replicated asynchronously to a second site. Create the storage group on a suitable SRP with SLO Platinum and workload type OLTP_REP, not OLTP: the _REP workload tunes for the write-heavy replication path, and the SLO controls service level for the whole group. If the database later needs a lower tier, the SLO change applies to every volume in that group, and compliance reporting restarts against the new target.

Traps

Assuming deduplication is always on; treating an SRP as a fixed RAID group; thinking SLO can be set per volume; forgetting that changing SLO or workload affects the entire storage group and its compliance reporting.