Abstract: What characteristics should an excellent self-developed database product have?
In the context of the continuous maturity of cloud computing technology, cloud databases have begun to rise, and are favored by small and medium-sized enterprises and Internet customers because of their excellent features such as on-demand expansion and payment on demand.
Although the database going to the cloud is inevitable, it is not a panacea.
Compared with self-built database, cloud database does not need to purchase and install any software and hardware, only pay for the service fee, take it as you go, reduce the input and cost of DBA.
However, in practical applications, the open source database cloud can only solve the requirements of small and medium-sized enterprises for simplified deployment, operation and maintenance, tuning, and extreme cost performance. It cannot meet the strict requirements of finance, government and enterprises on data security, response speed, reliability, and availability. Large and demanding enterprises.
When traditional database systems cannot cope with the storage requirements in the era of big data, enterprise customers urgently need new database products with dynamic capacity expansion, high throughput, and low cost.
In this context, it is imperative for cloud service vendors to develop their own cloud database products.
So, what characteristics should an excellent self-developed database product have?
With so many cloud databases, how to judge the pros and cons?
An excellent database product, or a database product that can solve the core business scenarios of enterprise-level users, must have the following characteristics:
1. It can support long-term massive R&D investment, guarantee its advanced nature, and can respond quickly to help users solve core problems;
2. Have a strong technical support team to meet business continuity requirements;
3. Possess a wealth of native support tools that can meet the basic demands of daily development, monitoring, operation and maintenance, and optimization;
4. Have a long-term development plan, a complete technical data and training certification system, and cooperate with key domestic ISV vendors and universities to create a long-term ecological strategy;
5. In extreme environments, it has a full-stack software and hardware integrated solution to ensure the continuity of users' business;
6. Joint innovation incubation with benchmark customers and strong ecological construction capabilities.
If the points mentioned above are to be achieved, it will not only test technical capabilities, but also ecological construction and service capabilities based on technology.
Combining the above characteristics, based on Huawei's internal IT and cloud bases' deep research and development and optimization capabilities for database cores for more than ten years, and understanding of customers' demands for high reliability and high performance, Huawei Cloud launched the enterprise-level distributed database GaussDB.
Technical breakthrough, product design wins
Let's first look at the design principles of Huawei Cloud native database, which generally include the following five aspects:
1. Separation of computing and storage, decoupling of master and slave;
2. The IO-intensive load is pushed down to the storage node to complete, such as redo processing and page reconstruction;
3. Make full use of the capabilities of cloud storage, the storage layer realizes independent fault tolerance and self-healing services, and shared access (single write and multiple read);
4. Give full play to the advantages of SSD, avoid write amplification caused by random write, reduce wear and delay; make full use of the random read performance of SSD;
5. When the performance bottleneck shifts from computing storage to the network, adopt new network technologies and hardware, such as RDMA.
Based on this, Huawei Cloud GaussDB has made many breakthroughs from the technical level.
Huawei Cloud GaussDB database is based on Huawei’s latest generation of DFV distributed storage, adopts a separate computing and storage architecture, and supports massive data storage. Both computing power and storage capacity can be elastically expanded and contracted in minutes; it also supports cross-AZ deployment with automatic data Backing up the local and OBS to achieve zero data loss can meet the high-performance, high-reliability, and high-availability requirements of large enterprises such as finance.
In other words, the GaussDB database not only solves some bottlenecks in open source databases on the cloud, but also optimizes some special services from the technical level.
Among them, it needs to be emphasized that the Huawei GaussDB database fully supports diversified computing power including Kunpeng and x86, and has E2E research and development capabilities from chips to servers, storage, operating systems, and databases, so it has database software and hardware performance tuning. The unique advantages of Huawei, such as the push-down storage of GaussDB database DB operator to achieve a 30% increase in performance compared to database partners, benefit from Huawei’s storage capabilities.
The unique software and hardware synergy advantages of the GaussDB database, and the support of leading benchmark customers such as ICBC, have enabled Huawei Cloud's self-developed database to quickly gain a foothold in the industry.
In addition to technical advantages, Huawei Cloud's database products have also made some detailed innovations in the design.
For example, in product design, it is universally applicable to the public, taking the usual path, allowing users to look simple and use it clearly. Especially in some harsh and complex scenarios, it is required to extreme simplification of professional skills, allowing a non-professional to complete business scenarios that are difficult for experts to complete through Huawei Cloud products.
For example, the current HUAWEI CLOUD DRS online migration scenario allows ordinary developers to complete the end-to-end online migration of the MySQL database through the DRS online migration tool.
Finally, by creating a development platform DAS that integrates development, operation and maintenance, and intelligent optimization, it greatly reduces the adaptation and learning costs caused by the diversity of database products, and makes use of all online database products.
Click to follow and learn about Huawei Cloud's fresh technology for the first time~