Brief introduction:At today’s event, DAS unveiled 8 new core autonomous features to address users’ core pain points, including 7×24 hour anomaly detection, automatic SQL flow limiting, automatic SQL optimization, automatic spatial optimization, automatic elastic scaling, intelligent parameter tuning, intelligent pressure test, real-time audit, etc. For the enterprise of SQL, database load, space problem, capacity assessment, security audit, from the abnormal findings, returning for positioning, and self-healing self-optimizing/from safety, can not only realize the quick stop, and will continuously optimize the database, without human intervention, let enterprises like experience “autonomous” using the database, Reduce database administration costs by 90%.
Ali a time box more: https://yqh.aliyun.com/live/cloudbox
In April 2020, Ali Cloud database autonomous service DAS was officially released, opening a new era of “automatic driving” of databases. After one year, the auxiliary autonomy mode of the database autonomy service DAS has supported 100% of the high availability instances of the whole network of Ali Cloud, and has more than 5000 customers, running in the autonomy mode of DAS, that is, authorizing DAS to carry out the self-repair, self-optimization, self-operation and self-security of the database.
Getting the best performance out of a database depends on a lot of database expertise. For most enterprises and application developers, it is still full of challenges, such as troubleshooting, quickly locating root causes and effectively stopping losses; In the process of rapid business iteration, continuous database tuning, SQL Review, etc., are very time consuming and require the ability of 7×24 hours on duty. Therefore, it is an inevitable trend for the database to move towards autonomy.
If the transformation from traditional database to cloud database is “car for carriage”, then DAS is adding an “automatic driving” engine to the “car” and has the ability of “automatic driving”.
At today’s event, DAS unveiled 8 new core autonomous features to address users’ core pain points, including 7×24 hour anomaly detection, automatic SQL flow limiting, automatic SQL optimization, automatic spatial optimization, automatic elastic scaling, intelligent parameter tuning, intelligent pressure test, real-time audit, etc. For the enterprise of SQL, database load, space problem, capacity assessment, security audit, from the abnormal findings, returning for positioning, and self-healing self-optimizing/from safety, can not only realize the quick stop, and will continuously optimize the database, without human intervention, let enterprises like experience “autonomous” using the database, Reduce database administration costs by 90%.
SQL problem – unattended way to achieve continuous database optimization
More than 70% of database performance problems are SQL problems, but the traditional solution below the lack of effective stop-loss means, and do not have the possibility of prevention in advance, continuous optimization.
DAS covers all parts of solving SQL problems through abnormal SQL location, automatic flow limiting of abnormal SQL, automatic SQL optimization and result verification, so as to realize timely detection, quick stop-loss, continuous optimization and verifiable.
Database load problem — to achieve automatic elastic expansion of the whole link to cope with the change of business load
For the vast majority of enterprises, the database is the backbone system, the database load soared, will cause huge losses to the business, but the warning method based on the threshold, there is a problem of large delay in abnormal discovery, often can not be in the load changes, rapid detection and timely use of stop-loss measures.
DAS’s 7×24 hour exception detection can detect database exceptions within 1 minute, and supports full link automatic elastic scaling (including bandwidth, Porxy, CPU/ memory /IO, nodes, etc.), which can quickly solve database load problems and ensure the continuous availability of business.
Spatial problems – database space problems self-repair, self-optimization
The spatial problems of database affect the performance and cost of database. DAS provides a complete self-repair and self-optimization scheme of “curing the symptoms” + “curing the root” for common spatial problems of database, which can save the cost of database use and improve the performance of database.
The full range of DAS autonomy enables enterprise users to maximize database performance
With the advent of the data era, each enterprise in the process of rapid business growth will inevitably encounter a variety of databases, data monitoring difficulties, high cost of operation and maintenance of common database operation pain points. Over the past year, DAS has established cooperation with more than 5000 enterprises to provide them with professional and efficient database management services. We take the well-known “AI+ education” listed company fluent English speaking and the well-known enterprise service provider Yonyou as examples, to show you how to solve the user performance pain points, reduce the cost of operation and maintenance as a database performance optimization tool.
Built database operation system and tools based on DAS. Based on the global workload, this system adopts real business flow and scene continuous optimization, completes the process loop of automatic tracking and rollback, realizes automatic SQL Review and automatic optimization, and makes the database complete true “autonomy”. In terms of operation process, the standardized non-inductive access and monitoring process of DAS can effectively avoid the possibility of artificial omission and misoperation in the operation process.
Due to DAS unified access and management, automatic diagnosis, automatic SQL Review and optimization features, the whole database operation and maintenance system is truly professional and efficient, greatly reducing the daily workload of fluent DBA speakers.
Sun Wenjie, Head of Fluency Talk Infrastructure, said, “The introduction of DAS can optimize the overall performance of Fluency Talk database, greatly improve the operation and maintenance personnel efficiency, and help the database management team to release manpower from heavy daily work and realize the transformation and upgrading of the team. DAS helps you use your database services smarter, more professionally, and more efficiently.”
Yonyou: DAS lightweight and personalized intelligent pressure test service helps Yonyou marketing cloud evaluate the database capacity and compatibility of MySQL database migration to PolarDB using real business scenarios and traffic.
In this process, DAS can capture real business traffic with low load, automatically generate pressure traffic through learning, support write traffic playback pressure test and automatic syntax conversion, all of which are helpful for various assessments to foresee possible risks, and provide multi-dimensional, high-perspective and all-round escort for database migration.
In terms of capacity assessment, the PolarDB specifications were accurately evaluated according to business requirements during the intelligent pressure test to avoid potential performance risks, and the specifications of migration instances were accurately evaluated to avoid unnecessary waste of IT resources.
Throughout the process, DAS has been escorting Yonyou’s database migration, identifying compatibility and performance risks in advance, and completing a reasonable capacity assessment. Finally, the business planning of database migration was successfully completed.
DAS sincerely invites you to enter the era of database “autopilot”
In just one year, DAS has gained the trust of more than 5000 customers and entered the era of “automatic driving” of the database. Apart from the demands of The Times, there are certain factors behind this success.
The complex business scenes inside Alibaba, more than 100,000 database instances, fast iterative business, super hot “Double 11”, etc., all provide DAS with a unique and irreplaceable “automatic driving” training ground, which helps DAS to exercise the ability of automatic driving. Although the DAS won’t be released until 2020, it has been in development since 2017. Inside Alibaba, DAS has experienced 3 years of continuous training, supported 3 Singles’ Day, automatically optimized more than 4900W of slow SQL, automatically recovered 4.6PB of space, and automatically optimized 12% of memory.
In addition, the transformation of cutting-edge research of Dharma Institute, the diversity of user business types, the complexity of database, and so on, all bring huge challenges to the automatic driving of database, many of which are difficult for professional DBAs, but DAS chose to rise to the difficulties, go ahead of The Times, and solve all the problems in advance. DAS, jointly developed by Aliyun and Damo Institute, has made great breakthroughs in anomaly detection, SQL diagnosis, root cause positioning and other core technologies. In the past 3 years, DAS related R&D team has published 4 papers in international top database conferences, including SIGMOD, VLDB Conference, etc.
In the past more than half a year, DAS, hundreds of billions of times the anomaly detection on a daily basis, at least more than 200000 times of SQL diagnosis and optimization, accumulated to help users in automatic processing of 10000 + fault, locate found 1 minutes, 5 minutes, 10 minutes stop-loss, without user manual intervention, maximum security database continuously available milestones. In the future, the database will be fully autonomous. DAS will continue to aim at the automatic driving of the database and continuously increase its autonomous capabilities of self-perception, self-decision, self-recovery, self-optimization and self-safety, so as to meet the deeper needs of users and the industry.
Related reading:
Technology revelation | large stimulating behind, how to effectively assess computing resources and planning database?
Blockbuster | database autonomous service will DAS papers to be included in the global top SIGMOD, pilot “database automated driving” a new era
| SQL request behavior recognition new functionality online to help solve the problem of finding a needle in a haystack in abnormal SQL detection
| DAS introduces the global Workload Optimization feature to automate SQL diagnostics
How does Autoscaling work in Aliyun
Copyright Notice:The content of this article is contributed by Aliyun real-name registered users, and the copyright belongs to the original author. Aliyun developer community does not own the copyright and does not bear the corresponding legal liability. For specific rules, please refer to User Service Agreement of Alibaba Cloud Developer Community and Guidance on Intellectual Property Protection of Alibaba Cloud Developer Community. If you find any suspected plagiarism in the community, fill in the infringement complaint form to report, once verified, the community will immediately delete the suspected infringing content.