Cloud Platform & Cloud Services: Storage as a Service
Cloud Platform & Cloud Services
This domain tests how storage is consumed as a cloud service and how Huawei Cloud storage services map to workloads. Know the three service models-block, file, and object-and the matching Huawei Cloud services: EVS for block volumes attached to ECS/BMS, SFS for shared file systems, and OBS for object storage. Also know backup and disaster recovery basics such as CBR and cross-region replication.
Core rules
- EVS volumes are block devices in an availability zone; they attach to one or more compute instances according to the volume type and sharing rules. Snapshots are incremental and AZ-scoped; they are not a full off-site backup by themselves.
- OBS stores objects in buckets with HTTP/S APIs; it is not mounted as a local block device. Use it for backups, archives, logs, and static content.
- SFS provides NFS/CIFS shared file access for multiple clients; it suits shared application files, not a single-node database requiring raw block latency.
- CBR provides backup policies for cloud resources; disaster recovery needs a defined RPO/RTO and a replication or backup target outside the failure domain.
- Cloud projects isolate resources and permissions; a subnet, security group, and route must allow the service traffic.
Worked example
A three-tier application needs a database, shared uploads, and long-term archives. Put the database on EVS because it needs low-latency block I/O; put shared uploads on SFS because multiple web servers need a common file namespace; put archives on OBS because cost and durability matter more than POSIX semantics. The reasoning is to match access semantics first, then performance and cost.
Traps
Do not mount OBS as a drive for a database; do not assume an EVS snapshot survives a region outage; do not ignore AZ placement; and do not treat a cloud backup policy as a tested recovery plan.