Best Practices for Public Cloud Deployment in China

Driven by the need for digital transformation, China's pace of cloud computing continues to accelerate. Gartner predicts that by 2024, 40% of China's IT infrastructure spending will be used for cloud services (see Figure 1).

Figure 1: China Cloud Infrastructure and Platform Services (CIPS) cloud migration rate

The growth of cloud services is largely driven by digital native companies, and for most Chinese companies, developing a cloud strategy is not an easy task, especially for hybrid cloud deployments involving private and multiple public cloud deployments, due to the complexity of China’s regulatory compliance and security requirements.

Organizations that have difficulty developing cloud strategies also have difficulty selecting candidate applications for public cloud migration. As a result, many cloud migrations end up using a basic “direct migration” approach rather than leveraging cloud-native services to complete the overall architecture transformation of existing applications.

In addition, cloud migration is difficult to carry out effectively because I&O leaders lack skills and experience in public cloud migration.

Three Phases of Building a Viable Path to Cloud Deployment

Most I&O leaders in China are aware of the importance of cloud computing, but many companies have not developed a specific cloud deployment plan. Effective cloud deployment planning can be divided into three stages: pre-implementation, implementation, and post-implementation. Pre-implementation activities can lay the foundation for the entire cloud strategic deployment plan. In the implementation stage, it is necessary to clarify the resources required for implementation, select vendors and services, and develop cloud workload configuration methods. In the post-implementation stage, activities related to future cloud management need to be carried out.

Conduct a detailed application portfolio assessment

Gartner believes that the deployment of cloud-native services is growing significantly in China. In China, cloud container services are used more frequently than other cloud-native services (e.g., functional platform as a service [fPaaS], serverless, event-driven). Choosing a container-native approach over a cloud-native approach helps migrate applications to the cloud and modernize all aspects of the infrastructure. However, over time, this approach may gradually become a barrier to transformation because the benefits of a large number of cloud-native services provided by cloud service providers (such as Amazon ECS/Lambda, Azure AKS/Key Vault, Alibaba Cloud ACS/PolarDB and other serverless services) cannot be fully utilized.

I&O leaders should start by engaging with their clients to conduct a detailed modernization assessment of their applications. Gartner identifies common drivers of application modernization based on business fit, business value, agility, cost, complexity and risk. Enterprises should determine the best migration approach for each workload, including increasing adoption of PaaS (especially for container-native services) to maximize the benefits of cloud native. Every organization will have its own answer to the question of how many applications a public cloud should host and the right way to migrate each workload to the public cloud (see Figure 2).

Figure 2: Evaluation of application portfolio deployed in China’s public cloud

Develop structured workload configuration standards

As I&O leaders determine the approach to application migration, they must also prepare the environment in which the applications will run. While some decisions are driven by factors outside the I&O department (such as vendor relationships and application fit), I&O must ensure that applications can run smoothly in the selected cloud environment. Mistakes in workload configuration selection can hinder the success of cloud deployments, so I&O leaders should work with architecture and application developers to develop workload configuration standards, including cloud service provider (CSP) selection, cloud reference architectures, resilience standards, and security frameworks.