Storage Systems and Technologies
Storage questions span media (HDD, SSD, NVMe), interfaces (SATA, SAS, NVMe over PCIe or fabric), and architectures (DAS, SAN, NAS, object). Know the tradeoffs: HDDs offer capacity per dollar, SSDs offer IOPS and low latency, and NVMe reduces protocol overhead versus SAS/SATA.
Core concepts: RAID levels and their failure tolerance - RAID 0 (striping, no redundancy), RAID 1 (mirror), RAID 5 (single parity, tolerates one drive failure), RAID 6 (double parity, tolerates two), RAID 10 (mirrored stripes). Understand rebuild risk: large drives mean long rebuild times and higher chance of a second failure. Know the difference between block (SAN), file (NAS), and object storage, and when each fits. Thin provisioning over-allocates capacity and can cause out-of-space events. Caching tiers and write-back caches improve performance but risk data loss on power failure unless protected.
Traps: assuming RAID is a backup (it is not - it does not protect against deletion, corruption, or site loss); confusing RAID 5 usable capacity with raw capacity; and forgetting that parity RAID has a write penalty.
Worked example / rule of thumb: six 10 TB drives in RAID 6 give 40 TB usable (4 data + 2 parity). Rule of thumb: usable capacity = (drives − parity drives) × drive size; and always keep RAID as availability, with separate backup for recovery.