ICT Infrastructure: Storage in the Data Center
ICT Infrastructure
This domain tests how storage fits into a data center: compute nodes, storage systems, and the SAN/LAN/IP network that connects them. For HCIA-Storage, know the difference between block, file, and object access; where SAN and NAS are used; and why redundancy and capacity planning determine availability.
Core rules
- Block storage (FC, iSCSI, NVMe over Fabrics) presents LUNs to servers; file storage (NFS, CIFS/SMB) presents shared directories; object storage uses HTTP/S and buckets.
- A Huawei OceanStor unified system can serve SAN and NAS, but the front-end ports, licensing, and access configuration must match the service.
- RAID protects against disk failure, not against deletion, corruption, or site loss. RAID 5 tolerates one disk failure; RAID 6 tolerates two; RAID 10 mirrors and stripes.
- Capacity planning separates raw capacity, RAID usable capacity, hot-spare capacity, and the file-system or LUN overhead.
Worked example
A project has eight 1.2 TB SAS disks. For RAID 6 with one global hot spare, usable raw-equivalent capacity is (8 - 2 parity - 1 spare) × 1.2 TB = 6.0 TB. If the design instead uses RAID 5 with one hot spare, it is (8 - 1 - 1) × 1.2 TB = 7.2 TB, but it tolerates only one disk failure. The reasoning is to subtract parity disks first, then any spare, and to state whether TB or TiB is meant.
Traps
Do not call RAID a backup; do not ignore hot-spare reservation; do not mix decimal TB with binary TiB; and do not assume every port can carry every protocol without configuration.