Scalability Design

The following figure shows scalability of Huawei Cloud products at each layer.

Figure 1 Scalability design example

The key points for designing scalability at each layer after cloud migration are as follows:

• Key points for scalability design at the application layer
  • If applications are deployed in containers using Huawei Cloud CCE, they can be scaled based on scaling policies of CCE. Application pods can be automatically scaled using AOM to efficiently handle workload fluctuations based on alarm policies.
  • If the application layer is deployed on ECSs, you can use Huawei Cloud Auto Scaling to implement horizontal scaling based on scaling policies.

• Key points for scalability design at the middleware layer
  • Message middleware layer: The underlying layer of Huawei Cloud DMS for RabbitMQ is a cluster. As the message volume and load increase, the specifications can be smoothly expanded.
  • Cache middleware layer: Huawei Cloud DCS Redis master/standby can seamlessly increase node specifications as the hot data capacity increases.

• Key points for scalability design at the data layer
  • Database middleware layer: Huawei Cloud DDM is deployed in a cluster. The DDM cluster specifications can be smoothly expanded to cope with increased database processing requirements.
  • Database layer: Huawei Cloud RDS can smoothly expand read replicas to handle a large number of reads. DDM is used to horizontally scale multiple instances. Large tables are horizontally split and evenly distributed to multiple database instances to improve the database capacity and performance. In addition, Huawei Cloud GaussDB adopts the decoupled storage-compute architecture, which supports scale-out within minutes and reduces interruption time.