Publication Date

2-28-2022

Document Type

Article

Department

Computer Engineering

Disciplines

Data Storage Systems | Systems and Communications | Systems Architecture

Publication Title

Big Data Research

Volume

27

DOI

10.1016/j.bdr.2021.100288

Abstract

The analytics in online performance monitoring systems have often been limited due to the query performance of large scale multidimensional data. In this paper, we introduce a faster query approach using the bit-sliced index (BSI). Our study covers multidimensional grouping and preference top-k queries with the BSI, algorithms design, time complexity evaluation, and the query time comparison on a real-time production performance monitoring system. Our research work extended the BSI algorithms to cover attributes filtering and multidimensional grouping. We evaluated the query time with the single attribute, multiple attributes, feature filtering, and multidimensional grouping. To compare with the existing prior arts, we made a benchmarking comparison with the bitmap indexing, sequential scan, and collection streaming grouping. In the result of our experiments with large scale production data, the proposed BSI approach outperforms the existing prior arts: 3 times faster than the bitmap indexing approach on single attribute top-k queries, 10 times faster than the collection stream approach on the multidimensional grouping. While comparing with the baseline sequential scan approach, our proposed algorithm BSI approach outperforms the sequential scan approach with a factor of 10 on multiple attributes queries and a factor of 100 on single attribute queries. In the previous research, we had evaluated the BSI time complexity and space complexity on simulation data with various distributions, this research work further studied, evaluated, and concluded the BSI approach query performance with real production data.

Keywords

Bit-sliced index, Bitmap index, Multidimensional data, Preference top-k queries, Performance

Comments

This is the Version of Record and can also be read online here.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.

Share

COinS