Performance Evaluation of Embedded Time Series Indexes Using Bitmaps, Partitioning, and Trees

Ould-Khessal, Nadir; Fazackerley, Scott; Lawrence, Ramon

doi:10.1007/978-3-031-17718-7_7

Cited by 2 publications

(4 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…They almost match the scenario with no index at all. The results are consistent with earlier experiments for SBITS (Ould-Khessal et al, 2022), showing that the number of I/Os to keep the index up to date is minimal. Perhaps the most surprising result is for the RadixSpline, because it needs to calculate spline points and update the radix table . Our benchmark shows that the overhead for these operations are small.…”

Section: Insertion Performancesupporting

confidence: 92%

“…The data sets cover a variety of sensor use cases including environmental monitoring, smart watches, GPS phone data, and chemical concentration monitoring. The environmental data sets sea and uwa were originally from (Zeinalipour-Yazti et al, 2005) and have been used in several further experimental comparisons (Fazackerley et al, 2021;Ould-Khessal et al, 2022). We also present the eCDF for the datasets in Figure 1.…”

Section: Resultsmentioning

confidence: 99%

“…The PGM has a slightly lower insert rate because some insertions trigger changes on multiple levels of the PGM, while RadixSpline triggers at most one change in the spline and radix table. The PGM insertion throughput remains competitive beating traditional embedded indexes such as B-Trees (Ould-Khessal et al, 2022).…”

Section: Insertion Performancementioning

confidence: 99%

“…SBITS also supports efficient querying by data value by using a user-customizable bitmap index. Prior experimentation (Ould-Khessal et al, 2022) demonstrated that SBITS outperforms MicroHash and other embedded indexing techniques.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Using Learned Indexes to Improve Time Series Indexing Performance on Embedded Sensor Devices

Ding¹,

Carvalho²,

Lawrence³

2023

Preprint

View full text Add to dashboard Cite

Efficiently querying data on embedded sensor and IoT devices is challenging given the very limited memory and CPU resources. With the increasing volumes of collected data, it is critical to process, filter, and manipulate data on the edge devices where it is collected to improve efficiency and reduce network transmissions. Existing embedded index structures do not adapt to the data distribution and characteristics. This paper demonstrates how applying learned indexes that develop space efficient summaries of the data can dramatically improve the query performance and predictability. Learned indexes based on linear approximations can reduce the query I/O by 50 to 90% and improve query throughput by a factor of 2 to 5, while only requiring a few kilobytes of RAM. Experimental results on a variety of time series data sets demonstrate the advantages of learned indexes that considerably improve over the state-of-the-art index algorithms.

show abstract

Section: Insertion Performancesupporting

confidence: 92%