Cloud Service Deployment, Migration and O&M
Domain 3: Cloud Service Deployment, Migration and O&M
This domain tests execution: landing a workload on Huawei Cloud, moving data and servers in, and running it with observability and governance. Know the migration tool boundaries and the O&M services that produce evidence.
Core rules
- Server Migration Service (SMS) migrates x86 physical or virtual servers to ECS; it is not a database replication tool.
- DRS handles database migration and synchronization; CDM handles data integration; OMS migrates object storage; DES moves large datasets offline when network transfer is too slow.
- Infrastructure as code: RFS and Terraform templates make deployments repeatable; CCE suits containerized workloads with auto-scaling.
- O&M: CES for metrics and alarms, LTS for logs, AOM for application and resource O&M, CTS for audit trails, and SMN for notifications.
- Tagging and enterprise projects are needed for cost allocation and quota tracking after migration.
Worked example
Migrate 60 TB of files and a 2 TB Oracle database over a 100 Mbps link. Pure network transfer of 60 TB at 100 Mbps is roughly 55–60 days, so DES is used for bulk files; OMS or CDM handles incremental object/file sync. DRS performs the database full load plus incremental sync, then a short cutover. SMS is used only for the application servers. The reasoning is tool fit and cutover window, not raw speed alone.
Traps
- Using DRS for file shares; it is database-oriented.
- Forgetting DNS TTL and connection strings during cutover, which extends the real outage.
- Deploying without tags and enterprise projects, making cost allocation and quota tracking fail later.