The database is the base of all enterprise business functions. R & D, testing, operation, and maintenance personnel in the enterprise need to operate or query the database every day, but they have different levels of skillsets with regards to using databases in a reliable manner. Thus database stability becomes a common challenge due to:
- New business application release, resulting in a lot of slow SQLs.
- Business promotions and insufficient database capacity estimation.
- Table structure or index design error.
- SQL or tables going live in GA without a thorough review.
Database troubleshooting and performance optimization have always been a challenging problem in the database field, even for professional DBAs. It might take them a long time to identify the root cause of a problem, mainly because:
- Difficult to obtain necessary information: problem diagnosis and performance optimization all depend on a large amount of system-generated data, including historical data.
- Difficult to analyze information. For someone who can provide an accurate analysis, it takes 1.many years of experience and 2. Knowledge of various usage scenarios. Experiences and scenarios are hard to scale because 1. not easy to pass on to another person, 2. They change constantly, and 3. not easy to consume.
- Difficult to optimize. Identifying the problem, knowing what to do, does not mean that the problem can be solved immediately. Some solutions require diving deep into the code optimization of the database engine layer.
With the popularity of cloud computing, enterprises can more easily use different databases according to different business scenarios or deploy databases in different environments. The challenges of managing multiple environments and multiple databases also follow:
- Professional DBAs who are proficient in multiple databases are rare talents and difficult to recruit;
- Difficulty in managing multiple databases deployed in multiple environments;
- Difficulty in depositing and inheriting management experience.
As the value of data increases, enterprises nowadays face more than ever internal or external attacks. Security risks such as data leak and data loss are also emerging rapidly.
Major global data breaches occur almost every month, and small leaks are constant.
- Unauthorized or unpredictable database access.
- Data leak.
- Data corruption.
- Hacker attack.
- Hardware and software bugs, causing data anomalies.
- Data loss due to misuse.
Database Autonomy Service (DAS) is a cloud service that implements database self-awareness, self-repair, self-optimization, self-maintenance, and self-security based on machine learning and expert experiences. DAS helps users eliminate the complexity of database management, and reduces service failures. DAS is key to effectively guarantee database stability, security and efficiency of database services on the Alibaba Cloud.
The service has been verified on all Alibaba Group databases for 2+ years until April 2020:
- Automatically optimized 40 million + SQL.
- Automatically reclaimed 4 PB of space.
- Automatically optimized 20% of memory.