I85PREP
⌘ K
Alibaba ACP Cloud Computing study materials
Alibaba ACP Cloud Computing study resources

Alibaba ACP Cloud Computing: Databases: ApsaraDB for RDS, PolarDB, Redis, MongoDB

Alibaba ACP Cloud Computing · Work through your exam, one topic at a time.

All guide sections

Databases: ApsaraDB for RDS, PolarDB, Redis, MongoDB

Relational options

ApsaraDB for RDS is managed MySQL, SQL Server, PostgreSQL, or MariaDB: automated backup, patching, and HA via primary/standby. PolarDB is a cloud-native relational engine with compute-storage separation, read replicas sharing storage, and fast scaling - better for high concurrency and elastic read scaling. Choose RDS for conventional managed relational needs; PolarDB for scale-out reads and rapid elasticity.

NoSQL options

ApsaraDB for Redis is an in-memory cache/key-value store with cluster and read/write splitting editions; used for caching, sessions, and leaderboards. ApsaraDB for MongoDB is document storage for flexible schemas.

HA and backup concepts

RDS and PolarDB offer multi-zone deployment (primary and standby in different zones). Backups include full and log backups enabling point-in-time recovery. Read replicas offload reads but are not automatic failover targets unless configured as such.

Traps

  • Assuming a read replica provides HA - it scales reads; failover depends on the standby/multi-zone setup.
  • Forgetting that connection endpoints differ (primary vs read-only endpoint) and that apps must target the right one.
  • Treating Redis as durable primary storage - it is a cache; persistence exists but data-loss windows remain.
  • Assuming PolarDB read replicas each have their own storage copy - they share storage, which is the key architectural difference from RDS.
  • Ignoring that parameter changes and major version upgrades may require a restart or maintenance window.

Rule of thumb

Pick RDS for standard managed relational with familiar engines; pick PolarDB when read scale-out, storage elasticity, or very high throughput matter. Use Redis in front of the database for hot reads, and always enable automated backups with a tested restore.

Worked example

A read-heavy product catalog: PolarDB primary with multiple read replicas behind a read-only endpoint, Redis caching the hottest queries, and daily backups plus log backup for point-in-time recovery.