HCIE-Cloud cheat sheet

Huawei Cloud IaaS core services: ECS (Elastic Cloud Server), EVS (Elastic Volume Service), VPC, ELB, OBS, IMS, and BMS; ECS instances are defined by flavor (vCPU/memory), image, disk type, and VPC/subnet placement.

ECS flavor families: general-purpose (s), memory-optimized (m), compute-optimized (c), GPU/accelerated (g/p), and high-performance (h); flavor naming convention is ., e.g. c6.large.2 means compute-optimized, generation 6, 2 vCPUs.

EVS disk types and typical IOPS ceilings: common I/O, high I/O, ultra-high I/O (SSD), and extreme SSD; disk performance scales with capacity, so a larger volume of the same type yields higher baseline IOPS and throughput.

VPC design rule: a VPC spans one region and can contain multiple subnets in different AZs; security groups are stateful instance-level filters while network ACLs are stateless subnet-level filters.

ELB types: dedicated load balancer (enhanced, supports layer 4/7, multiple AZs, higher performance) vs shared load balancer (basic, layer 4/7 with limited features); listeners define protocol/port and backend server groups define traffic distribution.

OBS storage classes: Standard, Infrequent Access, Archive, and Deep Archive; lifecycle rules transition objects between classes based on age, and bucket policies plus IAM policies jointly govern access.

AS (Auto Scaling) components: scaling configuration (template of flavor/image/disk), scaling group (bounds min/max/desired), and scaling policy (alarm-triggered, scheduled, or periodic) - the group enforces instance count within bounds.

Cloud service selection rule: use OBS for unstructured object data with HTTP access, EVS for block storage attached to a single ECS, and SFS for shared file storage mounted by multiple ECS instances concurrently.

Huawei Cloud management console layers: console (GUI), API, SDK, and CLI (KooCLI); IAM identities are accounts, users, user groups, and agencies - agencies are used for temporary cross-service or cross-account delegation without sharing permanent credentials.

IAM policy structure: Effect (Allow/Deny), Action, Resource, and Condition; explicit Deny overrides Allow, and permissions are the union of policies attached to the user and its groups.

Region/AZ model: a region is a geographic area with multiple isolated AZs; cross-AZ deployment provides HA within a region, while cross-region replication (e.g., OBS cross-region replication) provides DR across regions.

DR metrics: RPO is the maximum tolerable data loss (time), RTO is the maximum tolerable downtime; backup/restore gives high RTO, warm standby gives moderate RTO, active-active gives near-zero RTO.

Huawei Cloud DR services: CBR (Cloud Backup and Recovery) for backup/restore of ECS/EVS/SFS, SDRS (Storage Disaster Recovery Service) for cross-AZ/cross-region replication and failover of ECS and EVS, and VHA for application-level HA.

Monitoring and governance: CES (Cloud Eye) collects metrics and triggers alarms; CTS (Cloud Trace Service) records API operations for audit; LTS handles log collection and analysis.

Migration services: SMS (Server Migration Service) for host-level migration to ECS, DRS (Data Replication Service) for homogeneous/heterogeneous database migration, and OMS (Object Migration Service) for object storage migration.

Cost and elasticity rule: pay-per-use suits variable workloads, yearly/monthly subscription suits stable long-term workloads; AS combined with ELB provides horizontal elasticity, while flavor resize provides vertical scaling (requires restart).