Automatic Database Management System Tuning through Large-scale Machine Learning

Authors:

Dana Van Aken, Andrew Pavlo, Geoferry J. Gordon, Bohan Zhang

Motivation:

Database performance critically depends upon configuration knobs that control each aspect of the system, from the application's workload to the way the underlying hardware is utilized. Database tuning is complicated by the sheer number of configurable knobs, the dependencies between knobs, the number of possible values for each configuration and the difficulty in reusing previously determined configurations. The paper introduces Ottertune, an automated database tuning tool that recommends configurations based on previous experience. It compares the performance of OtterTune with that of Database-specific tuning tools and the recommendation of a human expert.

Pros:

1. Using only global knobs ensures that each that the metrics of one DBMS can also be used by another DBMS. Each tuning session can, therefore, have a larger repository of recorded metrics for training than traditional DBMS specific tuning tools.

2. Each tuning session has a repository of all the previously tested configurations and their effects available for learning. The information generated in each session can be used to refine the models and improve thereby improving the accuracy of the model over time

3. Pruning metrics using Factor Analysis and K-means clustering significantly reduces the search space of metrics to be considered for tuning and provides a concise representation of the workload

Cons:

1. The Lasso regression used in OtterTune makes use of polynomial terms to model the dependencies between the knobs. Depending upon the number of knobs being modeled, the number of higher-order interaction terms may grow explosively, affecting the complexity of the model.

2. Information about the effects of tuning a knob can only be gained through experience. Depending on the knob being tuned, the cost of conducting such a test may be extensive. However, without this prior information, it would be impossible to generate a learned tuning approach.

3. The tool is not completely automated since tasks such as blacklisting crucial knobs still require the intervention of a DBA. It also needs an external hint in order to discard useless metrics since it has no programmatic way of determining which metrics are truly useless.

4. It sacrifices fine-grained tuning at a tabular or component level by considering only global knobs in order to be able to map similar metrics across different databases.



-Karthik Unnikrishnan








Comments

  1. jon_weissmanFebruary 12, 2019 at 7:05 AM

    Excellent analysis. A deeper question is whether this approach can scale. How expensive is it to change configurations for each observation period? How many observation periods seem to be enough. Is there any hope of doing local knob optimization?

    ReplyDelete

Post a Comment

Popular posts from this blog

A Machine Learning Approach to Live Migration Modeling

Dynamic management of resources in data centers improves data center utilization in terms of energy consumption and device utilization. Virtual machines are among techniques to achieve this goal. To enable the VM to adapt to dynamic fluctuations of workloads, Live Migrated virtual machines are introduced, which stands for moving a VM from one physical host to another while the VM keeps running. Live migration enables load balancing, fault tolerance, ease maintenance, etc. However, live migration is costly in terms of data traffic. Having an understanding of performance metrics of different live migration algorithms is important so that one can rely on an efficient algrothm to migrate a running VM is important. Therefore, this work uses a Machine Learning (ML) approach to predict key metrics of different live migration algorithms, and then find the best suitable choice for live migration. Indeed, given resource usage as well as the characteristics of VM work load, this paper predict...
Read more

StormDroid: A Streaminglized Machine Learning-Based System for Detecting Android Malware

Introduction: The paper approaches the problem of Malware Detection using machine learning approach. In this paper, the authors propose considering new type of features along with two traditionally used set of features for developing techniques for malware detection. Along with this proposal, the paper makes other contributions as well. The authors provide a new framework which is able to process information from large data sets as streams, and efficiently perform malware detection. Motivation: Android is an open platform for development. It has certainly made easy for everyone to develop and publish apps.  The number of apps published on Google store is in millions. However, it has also made easier for hackers to reach out to a larger population for malicious attacks. Additionally, with surge in the number of attacks on devices, more data is available for analysis which can be used by the machine learning models to identify malicious apps. This is motivation to use machine ...
Read more

CrystalBall: Statically Analyzing Runtime Behavior via Deep Sequence Learning

Image
Author:  Stephen Zekany, Daniel Rings, Nathan Harada, Michael A. Laurenzano, Lingjia Tang, Jason Mars Intro: Understanding the runtime behavior of the software is critical in many aspects of program development. A conventional approach to this problem is using dynamic profiling, meaning we have to run the program multiple times with environments that representative enough for the actual working situation. Besides that, a dynamic profiling approach has other weakness such as cannot profiling a subset of functions or paths very economically and requires re-profiling each time any changes is applied to the code. This paper presents a statically profiling method, harnessing the ability of RNN to discover the hidden information of hot path in the code. Also, since the training process uses intermediate representation as input, the result will have better generalization power. Methodology: The dataset is in the form of a basic block feature; each sequence of basic blocks repr...
Read more