Cloud Solution Design Principles and the Huawei Cloud Service Portfolio
What to know
Every design question on this exam starts with the same discipline: translate a business requirement into a service choice, then justify it against five qualities - reliability, security, performance, elasticity, and cost. Huawei Cloud's global infrastructure is organised as regions (isolated, independent) containing availability zones (physically separate, low-latency, connected by redundant fibre). Nearly every service is region-scoped; many are AZ-scoped, which is the root of most HA design decisions.
Map the portfolio by category so you never reach for the wrong product:
- Compute: ECS (virtual machines), BMS (bare metal), CCE (managed Kubernetes), CCI (serverless containers), AS (auto scaling), DeH (dedicated host).
- Storage: EVS (block), OBS (object), SFS (file), CBR (backup).
- Network: VPC, ELB, EIP, NAT Gateway, Cloud DNS, Direct Connect, VPN.
- Security: IAM, KMS, WAF, HSS, Cloud Firewall, Anti-DDoS, CTS.
- Data: RDS, GaussDB, DCS, DWS, DDM, MRS, DLI.
- Operations: CES, LTS, AOM/APM, FunctionGraph.
Traps
Questions love to blur boundaries. DeH is a dedicated physical host inside the public cloud, not a private cloud. A backup is not a disaster recovery plan. A load balancer is not a firewall. Availability of a service or instance type can differ by region, so a textbook-perfect design may be infeasible in a given region. Also remember the shared-responsibility split: Huawei secures the platform, you secure your IAM, data, and guest OS configuration.
Rule of thumb
Adopt the highest abstraction that satisfies the requirement, then step down only for a named constraint. Stateless web tier → CCE or AS+ECS. Needs kernel control or a legacy ISV licence → ECS. Needs zero virtualization overhead, huge IOPS, or RDMA/IB → BMS. Hard compliance isolation → dedicated region or DeH.