期刊名称：Bulletin of the Technical Committee on Data Engineering
出版年度：2019
卷号：42
期号：2
页码：70-81
出版社：IEEE Computer Society
摘要：In big data era, database systems face three challenges. Firstly, the traditional empirical optimizationtechniques (e.g., cost estimation, join order selection, knob tuning) cannot meet the high-performancerequirement for large-scale data, various applications and diversified users. We need to design learningbasedtechniques to make database more intelligent. Secondly, many database applications require touse AI algorithms, e.g., image search in database. We can embed AI algorithms into database, utilizedatabase techniques to accelerate AI algorithms, and provide AI capability inside databases. Thirdly,traditional databases focus on using general hardware (e.g., CPU), but cannot fully utilize new hardware(e.g., ARM, GPU, AI chips). Moreover, besides relational model, we can utilize tensor model to accelerateAI operations. Thus, we need to design new techniques to make full use of new hardware.To address these challenges, we design an AI-native database. On one hand, we integrate AItechniques into databases to provide self-configuring, self-optimizing, self-monitoring, self-diagnosis,self-healing, self-assembling, and self-security capabilities. On the other hand, we enable databases toprovide AI capabilities using declarative languages in order to lower the barrier of using AI.In this paper, we introduce five levels of AI-native databases and provide several open challenges ofdesigning an AI-native database. We also take autonomous database knob tuning, deep reinforcementlearning based optimizer, machine-learning based cardinality estimation, and autonomous index/viewadvisor as examples to showcase the superiority of AI-native databases..
