Automatically Indexing Millions of Databases in Microsoft Azure SQL Database

Das, Sudipto; Grbic, Miroslav; Ilic, Igor; Jovandic, Isidora; Jovanovic, Andrija; Narasayya, Vivek; Radulovic, Miodrag; Stikic, Maja; Xu, Gaoxiang; Chaudhuri, Surajit

doi:10.1145/3299869.3314035

Cited by 49 publications

(23 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, PRP can help the optimizer to compare the costs of candidate plans or identify the query plan regressions with changed plans [15]. It may also significantly help an index tuner to detect performance regressions on candidate index configurations [14], which is a requirement for state-of-the-art index tuners [13].…”

Section: Enhanced Databasesmentioning

confidence: 99%

“…As an example, we simulate the production deployment of an ML enhanced index tuner. We use a state-of-the-art index tuner [13] to generate the plan (plan pair) space on 14 diverse database workloads (details in Section 7.2). We simulate a new deployment by holding-out the data in one database (i.e., the target database) and using the labeled data in the remaining 13 databases to train ML models for both ECP and PRP, resulting in 14 simulation runs per model in total.…”

Section: Prediction Error In Productionmentioning

confidence: 99%

“…We construct the target data for both ECP and PRP tasks by executing the queries from our workloads. We use a state-of-the-art index tuner [13] to recommend a set of indexes for each query, and enumerate subsets of the recommended indexes as the index configurations. The optimizer then generates candidate plans under each index configuration for a query.…”

Section: Evaluation Protocolmentioning

confidence: 99%

“…We use the optimizer's cost model as the crude cost estimator, which does not require any additional training data. To evaluate how the ADCP performs with different model retraining frequencies and budget sizes, we evenly split the total budget among multiple (3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15) iterations. In each iteration, we (1) use ADCP to label a batch of data points under the divided budget, (2) retrain the ML model with the existing and the newly labeled data, and (3) evaluate the model's prediction on the target data.…”

Section: Experiments Executionmentioning

confidence: 99%

“…To gather labeled data from production workloads, we leverage fork-offs (i.e., B-instances [13]) for databases in the cloud or backups for on-premise databases. These replicas can execute and instrument queries from production workloads to collect telemetry for data labeling.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Active Learning for ML Enhanced Database Systems

Ding

Das

et al. 2020

Proceedings of the 2020 ACM SIGMOD International Conference on Management of Data

Self Cite

View full text Add to dashboard Cite

Recent research has shown promising results by using machine learning (ML) techniques to improve the performance of database systems, e.g., in query optimization or index recommendation. However, in many production deployments, the ML models' performance degrades significantly when the test data diverges from the data used to train these models. In this paper, we address this performance degradation by using B-instances to collect additional data during deployment. We propose an active data collection platform, ADCP, that employs active learning (AL) to gather relevant data cost-effectively. We develop a novel AL technique, Holistic Active Learner (HAL), that robustly combines multiple noisy signals for data gathering in the context of database applications. HAL applies to various ML tasks, budget sizes, cost types, and budgeting interfaces for database applications. We evaluate ADCP on both industry-standard benchmarks and real customer workloads. Our evaluation shows that, compared with other baselines, our technique improves ML models' prediction performance by up to 2× with the same cost budget. In particular, on production workloads, our technique reduces the prediction error of ML models by 75% using about 100 additionally collected queries.

show abstract

Section: Enhanced Databasesmentioning

confidence: 99%

Section: Prediction Error In Productionmentioning

confidence: 99%

Section: Evaluation Protocolmentioning

confidence: 99%

Section: Experiments Executionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Active Learning for ML Enhanced Database Systems

Ding

Das

et al. 2020

Proceedings of the 2020 ACM SIGMOD International Conference on Management of Data

Self Cite

View full text Add to dashboard Cite

show abstract

Internet Information Processing Technology Based on Big Data Analysis

Zhao¹,

Zhang²

2023

Lecture Notes on Data Engineering and Communications Technologies

View full text Add to dashboard Cite

$$\hbox {CDBTune}^{+}$$: An efficient deep reinforcement learning-based automatic cloud database tuning system

Zhang

Zhou

et al. 2021

The VLDB Journal

View full text Add to dashboard Cite

Configuration tuning is vital to optimize the performance of a database management system (DBMS). It becomes more tedious and urgent for cloud databases (CDB) due to diverse database instances and query workloads, which make the job of a database administrator (DBA) very difficult. Existing solutions for automatic DBMS configuration tuning have several limitations. Firstly, they adopt a pipelined learning model but cannot optimize the overall performance in an end-to-end manner. Secondly, they rely on large-scale high-quality training samples which are hard to obtain. Thirdly, existing approaches cannot recommend reasonable configurations for a large number of knobs to tune whose potential values live in such high-dimensional continuous space. Lastly, in cloud environments, existing approaches can hardly cope with the changes of hardware configurations and workloads, and have poor adaptability. To address these challenges, we design an end-to-end automatic CDB tuning system, $${\texttt {CDBTune}}^{+}$$ CDBTune + , using deep reinforcement learning (RL). $${\texttt {CDBTune}}^{+}$$ CDBTune + utilizes the deep deterministic policy gradient method to find the optimal configurations in a high-dimensional continuous space. $${\texttt {CDBTune}}^{+}$$ CDBTune + adopts a trial-and-error strategy to learn knob settings with a limited number of samples to accomplish the initial training, which alleviates the necessity of collecting a massive amount of high-quality samples. $${\texttt {CDBTune}}^{+}$$ CDBTune + adopts the reward-feedback mechanism in RL instead of traditional regression, which enables end-to-end learning and accelerates the convergence speed of our model and improves the efficiency of online tuning. Besides, we propose effective techniques to improve the training and tuning efficiency of $${\texttt {CDBTune}}^{+}$$ CDBTune + for practical usage in a cloud environment. We conducted extensive experiments under 7 different workloads on real cloud databases to evaluate $${\texttt {CDBTune}}^{+}$$ CDBTune + . Experimental results showed that $${\texttt {CDBTune}}^{+}$$ CDBTune + adapts well to a new hardware environment or workload, and significantly outperformed the state-of-the-art tuning tools and DBA experts.

show abstract

Automatically Indexing Millions of Databases in Microsoft Azure SQL Database

Cited by 49 publications

References 16 publications

Active Learning for ML Enhanced Database Systems

Active Learning for ML Enhanced Database Systems

Internet Information Processing Technology Based on Big Data Analysis

$$\hbox {CDBTune}^{+}$$: An efficient deep reinforcement learning-based automatic cloud database tuning system

Contact Info

Product

Resources

About