Shulou(Shulou.com)05/31 Report--

Many beginners are not very clear about how to solve the pain points in the transformation of Internet architecture in database practice. in order to help you solve this problem, the editor below will explain it in detail. People with this need can come and learn. I hope you can get something.

Pain points in the Development of ▍ Internet Database

In the process of the development of Internet database, we will encounter some pain points when we do IT construction. In this process, how to solve these pain points is very important. *, many enterprises will face the process of data and business from simple to complex, and the amount of data from small to large. In this process, the data may not be stored in a single database. So how do we improve scalability at this time? Second, our business has developed, and so has our data, but have we kept up with the pace of business development? Third, our enterprise will have some core data, such as customer user information, including some basic configuration information, in the process from small to large. This information is often used by many departments in the enterprise, so how to share the core data? At the same time, in the process of enterprises using different databases, how to exchange data between these databases is also a pain point we are facing. Fourth, our R & D personnel and control personnel will continue to increase with the growth of business scale. How can we ensure the security and standardization of data when there are many people without affecting the efficiency of development and testing? data is the core asset of an enterprise. How to improve the disaster recovery capability of the database to ensure that our data will not be lost as a result of natural and man-made disasters? The following is to introduce our products with the pain points encountered in the whole transformation process of Alibaba.

The picture above shows the development of Alibaba's distributed database. At present, Alibaba solves the problem through distributed database. DRDS products support multi-dimensional split, for example, we can split according to the ID of the transaction, the customer's ID, and the time.

The picture above shows a typical customer case of distributed database DRDS and a case of telecom operator. It has the pain points of rapid growth and scalability bottleneck in the process of business development. At the same time, if the customer uses the traditional method to split the database, it will face the problem of high cost of application and transformation. DRDS improves storage expansion and computing power, and transparent access application, smooth expansion, transparent generation and configuration, and transparent read-write separation solve the problem of high cost of customer application transformation.

▍ Alibaba data transmission

The product has three core capabilities: support for transport synchronization, support for database migration, and support for data subscription. Data changes can be converted into messages sent downstream, which can be used for phone consumption based on this message.

The image above shows the architecture of the data transfer DTS. In the lower layer, we support cloud databases, including ECS self-built databases, user IDC databases, as well as groups and ants. Data can be migrated to different destinations through the product, and data synchronization subscriptions are also supported. We can convert the data from the database into files.

The figure above shows a data transfer DTS customer case. The main problems faced by customers are: the local Oracle database is the performance bottleneck of the customer's business, and when active, the stored procedures on Oracle occupy nearly 80% of the CPU of the instance. It also requires that the business on Oracle be migrated to Aliyun DRDS without downtime and running continuously. The solution is to build a real-time synchronization link between Oracle and DRDS instances through DTS to ensure real-time synchronization of DRDS and Oracle data. After the data between DRDS and Oracle instances are synchronized in real time, the business chooses to switch part of the traffic to DRDS on Ali Cloud at any time. Through DTS, the business traffic generated on DTS is flowed back to Oracle,Oracle and DRDS to provide disaster recovery for each other during the completion of business cutover to ensure business continuity. Alibaba's past life and present life of data management. Through the information-based unified management of the database, and through the regularization, automation, rules and smooth self-help changes of non-operation and maintenance operations, DMS has been officially commercialized in the public cloud and also supports privatized output.

▍ hybrid cloud database management HDM

The figure above shows the hybrid cloud database management HDM architecture and capabilities. Our daily operation and maintenance work can basically be solved by HDM. The core competencies of HDM include: hybrid cloud database architecture, unified multi-environment management, automatic operation and maintenance, low cost, disaster recovery switching, rapid elastic capacity expansion, one-click switching, stability, disaster recovery, and flexibility.

The figure above shows a disaster recovery solution based on HDM, DBS, and DTS. The * * level way is that we throw the local backup directly to the OSS by pulling files. Of course, this is relatively low-level. The second is that we can upload real-time backup to the cloud through DBS, and then restore it locally or on RDS through DBS after backup. Third, we synchronize the database under the cloud to the RDS on the cloud through DTS. The RDS on our cloud forms a disaster recovery with that under the cloud, that is, once our local IDC cannot be accessed, we can access the RDS on the cloud to ensure the continuity of the business. This scenario means that we can achieve two-way synchronization under and on the cloud. This two-way synchronization can actually achieve multi-activity in different places, that is, we have a database in computer room An and a database in computer room B for two-way synchronization. These two-way synchronization can support IDC access and can also be accessed on the cloud. As long as the application side is built well enough, then this dual work can be achieved. At the same time, based on our HDM, we can get through our user IDC, cloud on and off cloud, and realize the one-click switch of database disaster recovery.

▍ Enterprise Database Product Architecture

I just shared a few products with you. Let's review some of the capabilities of these products. The first is the distributed database DRDS. We can understand that it is a storage layer to meet the scalability of the database. The second is our data transmission. Data transfer realizes the ability of data migration, synchronization and subscription, which can connect the database outside the cloud with the database on the cloud. At the same time, it is not limited to these functions, including our own internal. The third is our data management, it can be understood that it is a database-level DevOps, so that developers and testers in our enterprise can complete data management without much interaction. There is our database backup DBS to meet the needs of enterprises to back up data at a lower cost. It is our hybrid cloud data management HDM, which can access the database on the cloud and under the cloud, and achieve one-click disaster recovery switch.

▍ Enterprise Database solution

In addition to the product, we will also output some solutions, the solution is dependent on the product, to provide some solutions for our enterprises. Like R & D efficiency assurance and data security solutions, this is based on DMS. In other words, when we export, we will help enterprises use our products and give our enterprises some advice, such as how to achieve a more efficient, safer and more compliant data management system. The second is DTS. DTS itself has the functions of data migration, synchronization and subscription, so we will help enterprises to use DTS together and be able to move data to the cloud. Then we will output multi-level disaster recovery solutions to our enterprises based on HDM, DBS and DTS.

Is it helpful for you to read the above content? If you want to know more about the relevant knowledge or read more related articles, please follow the industry information channel, thank you for your support.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.