Huawei Cloud Server Migration Service (SMS) is specifically designed for migrating source servers from other platforms to Huawei Cloud. It supports continuous synchronization, allowing for a phased migration where workloads can be replicated to the cloud while the source remains online, enabling a low-downtime or even zero-downtime cutover. CBR is for backup, ECS is a compute resource, and VPC is for networking; none are the primary migration tool.

Cross-AZ high availability is a core design principle for disaster recovery at the AZ level. Deploying application instances across multiple AZs and using ELB to distribute traffic ensures that if one AZ fails, the ELB can route traffic to the healthy instances in the other AZ(s). Options A, C, and D describe configurations confined to a single AZ, which does not protect against an AZ-wide failure.

The combination of Cloud Eye (for monitoring and alarms) and Auto Scaling (for executing the scaling actions) is the native, managed service approach for implementing elastic scaling. It is operationally efficient, automated, and follows best practices. Manual intervention (A), custom scripts (B), and static over-provisioning (D) are either error-prone, not scalable, or cost-ineffective.

Spot Instances offer spare compute capacity at a significant discount (up to 90% off On-Demand prices). For interruptible, fault-tolerant, and short-duration batch processing jobs like a monthly report, they are ideal for maximizing cost savings. On-demand (A) is flexible but more expensive for a 48-hour run. Subscriptions (B) are cost-effective for steady-state, always-on workloads, not temporary ones. EVS (D) is storage, not compute billing.

While all options can store data, Huawei Cloud Key Management Service (KMS) is the dedicated service for creating, managing, and using cryptographic keys securely. It is the correct choice for managing the encryption keys needed to protect sensitive data like passwords, which can then be stored securely elsewhere (e.g., in a configuration management tool or database). Storing plaintext (A) is insecure, OBS (B) is for general object storage, and Redis (C) is a caching database, not a dedicated key management system.

Using a bastion host (or jump server) is a standard security practice. The application servers reside in a private subnet with no public IP, isolating them from direct internet access. Administrators first connect to the hardened bastion host in a public subnet, and from there, connect to the private instances. This creates a single, auditable choke point. Options A and D still expose management ports directly, albeit with restrictions. A NAT Gateway (B) handles outbound traffic, not inbound administrative access.

ECS Auto Recovery is a specific feature that monitors the underlying physical server health. If a hardware failure is detected, it automatically migrates the affected ECS instance to a healthy physical server, minimizing application downtime. Cloud Eye (B) can alert you, but doesn’t perform the recovery action itself. Anti-DDoS (A) protects against attacks, and an EIP (C) is a static public IP address; neither directly addresses hardware failure recovery.

Scalable File Service (SFS) is a fully managed, shared file storage service that provides a standard file system interface (NFS), supports POSIX semantics, and allows concurrent access from multiple ECS instances. It scales automatically, making it ideal for analytics, media processing, and shared content. OBS (B) is object storage with a different access model (HTTP/REST). EVS (C) is block storage typically attached to a single instance. Manual sync (D) is not a managed, concurrent access solution.

For demanding database workloads (like OLTP), the priority is on consistent low-latency and high IOPS. Low-latency EVS disks (backed by NVMe SSDs) are specifically designed for this scenario, offering the highest performance. Pairing them with a Compute-optimized or Memory-optimized ECS (which often have better CPU-to-memory ratios and network performance) is a common best practice. Ultra High I/O (A) is a legacy tier, Common disks (B) are HDD-based and too slow, and High I/O SSDs (C) are good but not the highest tier for extreme performance needs.

Huawei Cloud Server Migration Service (SMS) is specifically designed for migrating source servers from other platforms to Huawei Cloud. It supports continuous synchronization, allowing for a phased migration where workloads can be replicated to the cloud while the source remains online, enabling a low-downtime or even zero-downtime cutover. CBR is for backup, ECS is a compute resource, and VPC is for networking; none are the primary migration tool.0