Cloud Native Tech Summit was recently held in Thailand. More than 200 leaders and experts from many sectors and large enterprises in Thailand were invited to the summit to discuss Thailand's digital transformation. Zhang Hu, Solution Director of Huawei Cloud Database Service Product Dept, attended the salon and shared Huawei Cloud's innovative technologies and best practices for using cloud-native databases. Huawei Cloud is helping Thailand enterprises unleash digital productivity for service growth.

Cloud-native databases are becoming the preferred choice for enterprise customers

Mr. Zhang said that the fast development of Internet technologies has been pushing enterprises to create many new business models, such as e-commerce, Internet games, digital banking, and advanced logistics. New applications and scenarios bring new challenges. Traditional databases are unable to meet these new requirements. They lack the performance, high availability, support for hybrid loads, and cost-effectiveness that today's enterprises need.

To keep up with customer demands, database architectures have evolved through three distinctive phases: on-premises, cloud-based, and cloud-native.

• In the on-premises phase, customer databases are deployed in the enterprises' data centers. Software and hardware specifications are fixed for a certain period of time. Database capacity is reserved for three years, which is a waste of resources in the initial stages.
• The cloud-based deployment phase can be divided into two stages. First, customers deploy databases using IaaS and they design enterprise-grade high availability, high reliability, and O&M features by themselves. In the next stage, they use cloud database services provided by cloud service providers or managed open-source databases. These providers provide service installation and upgrade and many advanced features, such as primary/standby HA components, automated backup, security audit, and visualized monitoring and alarm reporting, which makes the database O&M easier.
• In the cloud-native phase, we have cloud-native databases. Database vendors bring high performance, scalable, reliable, easy-to-use, intelligent, and cost-effective databases to market. The databases are designed based on the virtualization and pooling technologies of cloud platforms and have a decoupled architecture that separates compute from storage. They also integrate many features to guarantee database reliability, such as log as data, operator pushdown, multiple replicas, and cross-AZ/region DR. Cloud-native databases have become the preferred choice for more and more customers migrating services to the cloud.

Huawei Cloud creates application-centric cloud-native databases with three major characteristics. They are serverless, regionless, and modeless

As digital transformation deepens, enterprise applications have changed from being resource-centric to application-centric, from being region-centric to traffic-centric, and from being load-centric to data-centric. To meet new enterprise requirements, Huawei Cloud provides cloud-native databases with re-architected kernels and architectures. Huawei Cloud has also proposed three major aspects to focus on:

• Serverless: superior scalability

If a failure occurs or an instance class changes, the time needed for scheduling resources is not measured in minutes. It can be done in just seconds, and user experience is never impacted. Huawei Cloud GaussDB(for MySQL) is a full-stack serverless database service, and it can be billed based on new computing units, greatly improving resource utilization.

• Regionless: data available in all regions

Services are accessible anywhere and cross-region HA is achieved. By accessing data from the nearest region, you get faster responses.

• Modeless: smooth user experience

A single portal is provided, where you can enjoy automated selection of an appropriate engine to process service requests and automate conversion between different data models. Using a unified portal simplifies data processing and management.

Huawei Cloud's relational database GaussDB(for MySQL) and non-relational database GaussDB NoSQL have great advantages over similar open-source databases. For example, GaussDB(for MySQL) is 100% compatible with MySQL 8.0 and provides up to 7x the performance of the open source version. GaussDB(for MySQL) supports up to 128 TB of storage and millions of QPS. Cross-region DR can be performed in real time. GaussDB NoSQL supports Redis, InfluxDB, and Cassandra engines, providing high scalability and performance, rock-solid reliability, massive storage, and enhanced DR capabilities. GaussDB(for Redis) provides up to 2 Gbit/s of bandwidth, tens of millions of QPS, and sub-millisecond latency. It can be scaled in seconds with a lightweight, serverless architecture and supports three-AZ deployment and up to 36 TB data storage. It also provides powerful enterprise-grade features such as multi-tenant management by tenant and point-in-time recovery (PITR).

Huawei Cloud's cloud-native databases, tested in nearly every industry, are becoming increasingly mature

With unique advantages and leading technologies, cloud-native database services of Huawei Cloud have been widely adopted in industries such as Internet, e-commerce, gaming, and automobile manufacturing. Huawei Cloud databases help enterprises achieve agile development with extensive use cases.

For example, in the Huawei Mobile Service (HMS) project, GaussDB(for MySQL), GaussDB(for Cassandra), and GaussDB have helped HMS grow into the stable and efficient service platform that it is today. So far, more than 3,000 instances, 16,000 nodes, and 11 petabytes of data have been rolled out. These databases help HMS handle trillions of records and handle millions of QPS. The overall resource usage and O&M efficiency have been improved by 30%.

Cloud native technology is being adopted for more and more critical industry scenarios. Huawei Cloud databases will continue to evolve towards serverless, regionless, and modeless technologies and to help provide intelligent databases for core services.