Smart Time Series Data Compression
Get 99% accuracy using only 1% of your data
700 times faster analytical results with 99% accuracy on a 6.3TB data set
Making time series data smaller, faster and more valuable
Datalytyx Gallium is a machine learning algorithm that massively reduces the amount of IoT/time series data needed to be transported, stored or processed to achieve results with guaranteed accuracy.
Explore TBs of Data
Query huge data sets in seconds instead of hours
99% Accuracy in a fraction of the time
Gallium reduces the amount of data you use for key decision making to 1% and maintains a guaranteed accuracy rate of 99%. Get 700x faster analytics with the same data warehouse/BI tool. Reports that used to take hours to generate can be done in seconds. Huge volumes of data that used to be too large to interrogate can now be explored in seconds. It’s simple to integrate into your current ETL/ELT pipeline.
Gallium’s machine learning algorithm is more than 50 times as accurate as data averaging or data sampling. By intelligently classifying IoT data, Gallium massively reduces the volumes travelling over your link. Get hundreds more metrics over the same limited bandwidth while dynamically adapting to local events and link capacity.
Get more value with less data
It’s rare to need 100% accuracy from your data to make a decision. Most decisions can be made as long as you know your data is guaranteed to be almost completely accurate.
The graphs below show Averaging (the industry defacto way to reduce time series data volumes), Sampling and Gallium, all at a 100:1 data reduction.
"...a lot of machine-generated data has a poor signal-to-noise ratio. It contains valuable data but also a lot of 'noise'. Therefore, you often must store all of it to find the valuable bits."
– Cloud Data Warehousing for Dummies by Snowflake