I85PREP
⌘ K
010-151 DCTECH study materials
010-151 DCTECH study resources

010-151 DCTECH: Data Center Server and Storage Architecture

010-151 DCTECH · Work through your exam, one topic at a time.

All guide sections

Data Center Server and Storage Architecture

This domain tests how Cisco UCS components are cabled and addressed, how RAID and disk media behave, and how SAN and service-profile abstractions connect servers to storage. Expect capacity arithmetic, identity questions, and protocol matching.

UCS hardware model

  • A B-series chassis holds 8 half-width blades (or 4 full-width), two IOMs, four PSUs, and fan modules; the IOMs uplink to Fabric Interconnects.
  • C-series rack servers use a standalone CIMC or are integrated into UCS Manager.
  • VICs present vNICs and vHBAs; mezzanine adapters do the same for blades.
  • Service profiles and pools (MAC, WWNN/WWPN, UUID, boot, BIOS, firmware) make servers stateless. A replacement blade inherits the profile and its identity.

RAID and media

  • RAID 0 stripes with no protection; RAID 1 mirrors; RAID 5 stripes with one parity disk (one failure); RAID 6 dual parity (two failures); RAID 10 mirrors stripes for write performance.
  • Write penalty: RAID 1/10 = 2, RAID 5 = 4, RAID 6 = 6 I/Os per host write.
  • SAS is dual-port and enterprise-grade; SATA is single-port and cheaper; NVMe over PCIe has the lowest latency; SSDs shift the discussion to endurance and wear.
  • A hot spare is idle capacity that rebuilds automatically and is not counted as usable.

Storage protocols and SAN

  • Fibre Channel uses WWPNs and zoning on MDS switches; VSANs separate fabrics logically.
  • FCoE carries FC frames over lossless 10G Ethernet with DCB/PFC; NPV/NPIV let the Fabric Interconnect act as an access switch to an upstream SAN.
  • iSCSI and NFS run over IP; LUN masking controls which initiator sees a LUN.

Worked example

A RAID group has six 900 GB drives and one global hot spare. The other five drives are RAID 5. Usable capacity = (5 - 1) x 900 GB = 3,600 GB. If those five data drives were RAID 10, usable would be 2 x 900 GB = 1,800 GB, and RAID 10 needs an even drive count. The hot spare is excluded from both calculations. Reasoning: subtract parity first, multiply by drive size, never add the spare.

Traps

  • Confusing WWNN (node) with WWPN (port).
  • Assuming RAID protects against a lost array or site failure.
  • Counting a hot spare as usable capacity.
  • Forgetting the RAID 5 write penalty is 4, not 1.