Here is a quick guide to help you choose a HUAWEI CLOUD database suitable for your services.

Relational Databases

You can choose a relational database if your services are in the following categories:

Websites: In read-intensive scenarios, RDS for MySQL enables you to add read replicas to expand the read capability of your database. By working with Distributed Database Middleware (DDM), RDS for MySQL delivers automatic read/write splitting and read load balancing.

Mobile applications: RDS for PostgreSQL empowers location-based applications to store and compute geo-location data. You can use RDS for PostgreSQL together with DDM to split databases and tables.

Gaming: To deal with explosive growth of data storage and read/write requests, you can use HUAWEI CLOUD RDS to quickly expand the storage capacity, change specifications, or deploy new game partition databases. To archive game data or roll it back to any point in time, you can use RDS automated backups and point-in-time restore (PITR).

E-commerce: RDS DB instances with high specifications can easily handle high concurrent requests during service spikes. For e-commerce services with high requirements on service continuity, you can deploy HUAWEI CLOUD RDS in hot standby mode across AZs to achieve high availability.

Finance: Financial services always attach great importance to service continuity and data reliability. The distributed GaussDB(for MySQL) service caters to these needs in compliance with financial security requirements. By working with Database Security Service (DBSS), GaussDB can monitor and intercept SQL injections in real time, prevent data breaches, and audit database logs.

Non-Relational Databases

NoSQL databases sacrifice data consistency in exchange for high performance and ease of use. Common applications for non-relational databases include:

Enterprise system log storage: Enterprise application logs can be stored in databases as documents. Each document is a self-contained data unit and a collection of data items. Document databases do not require fixed schemas and can store different types of information.

Recommended product: DDS

User data storage: Key-value databases are great for storing user data such as sessions, user profiles, and shopping cart data in e-commerce scenarios, as well as roles, character scores, and gadgets acquired in gaming scenarios. This sort of user data is usually associated with IDs (keys) and data can be retrieved by referencing its key.

Recommended product: GeminiDB

Column families for logs and blogs: Log data is stored in different columns, and each application can write information to its own column family. For example, tags, categories, and articles can be stored in different column families. Column store databases store data in columns instead of rows and load, search, and aggregate data faster.

Recommended product: GeminiDB

Time series data analysis: Time series data are simply measurements or events that are tracked, monitored, downsampled, and aggregated over time. This could be IoT sensor data, trades in a market, real-time monitoring of hardware and software, or environmental data analytics. Time series databases support fast writing, persistency, and multi-dimensional aggregation query of time series data. Each record is time-stamped for efficient analysis.

Recommended product: GeminiDB

If your service has a large amount of structured data and transactional operations, then a MySQL database is preferred. If service spikes are common, use non-relational databases such as MongoDB. In brief, selecting the right database is vital for your business development. Always take your service characteristics and requirements into consideration when selecting a database.