Quasar is the universal adapter between raw timeseries data and Automated Intelligence (AI).
Quasar radically simplifies data engineering by delivering a fully integrated data platform for timeseries data, from ingestion to feature extraction.
Quasar captures, normalizes, and stores an unlimited volume of events in the most efficient way possible.
Events are organized as timeseries that can be queried using SQL or through any of the numerous API and connectors Quasar offers.
Features can be extracted on demand using SQL or through the numerous API and connectors Quasar offers.
Timeseries data is typically large as it is cumulative.
When data volumes increase, there are two main approaches: the data lake approach and the data warehouse approach.
Both approaches have inherent limitations and ask you to make a tradeoff between ingestion and querying.
This makes it very difficult to connect automated decision making to your business.
Thanks to its radical design, Quasar goes beyond these tradeoffs and delivers unlimited capture, at low latency, while offering real time feature extraction and querying capabilities.
The core principle of Quasar is to enable you to capture everything you want in raw format and offer you the possibility to transform it however you want when you need it. The idea is that crucial information may be hidden in weak signals, and you never know in advance which part of the data you need.
Build a true digital twin of your business environment!
Not only you can’t know which raw data you need, but it’s also impossible to predict which features will have to be extracted. Storing every possible feature is no longer possible when the raw data volumes get close to the petabyte, and using dedicated programs is neither flexible nor cost-effective.
Quasar lets you work on raw data and extract features on the fly so that data is ready to be used by your models and tools.
Quasar's brain is QuasarDB, a high-performance, distributed, column-oriented timeseries database management system designed from the ground up to deliver real-time on petascale use cases.
Up to 20X less disk usage. Quasardb ingestion and compression capabilities are unmatched.
Up to 10,000X faster feature extraction. QuasarDB can extract features in real-time from the raw data, thanks to the combination of a built-in map/reduce query engine, an aggregation engine that leverages SIMD from modern CPUs, and stochastic indexes that use virtually no disk space.
Learn more about QuasarDB in our technological white paper.
The devil is in the details. Models are sensitive to subtle variation, and crucial information may be hidden in weak signals, and you never know in advance which part of the data you need.
That’s why Quasar enables you to capture everything you want in raw format and offer you the possibility to transform it however you want when you need it.
Not only you can't know which raw data you need, but it's also impossible to predict which features will have to be extracted. Storing every possible feature becomes more and more expensive as data grows, and using dedicated programs is neither flexible nor cost-effective.
Quasar lets you work on raw data and extract features on the fly so that data is ready to be used by your models and tools.
Quasar has a unique data model based on tags that gives you both raw speed and extreme flexibility.
Timeseries are organized in tables. Tables are so inexpensive that you can have millions of them. Use SQL to query them.
Tag your tables, query based on tags. Or tables. Or both. Possibilities are endless.
Tags can be changed at any time and can be recursive. Reorganizing data is instant and effortless.
Quasar write capabilities scale linearly with the number of nodes in the cluster.
Quasar highly optimizes your storage capacity by combining extremely efficient compression and optimized I/O.
Quasar saves you the trouble of writing custom programs by supporting a wide range of built-in feature extraction capabilities, greatly increasing the speed at which you can work on raw data.
Quasar can extract features from raw timeseries data on the fly, this goes from simple averaged to more complicated slope and kurtosis computations, combining that with grouping and pivoting.
Quasar has a dashboard framework that allows for the construction of bespoke web dashboards.
The Quasar framework is perfect for a project that has precise requirements in terms of visualization and reporting.
For users already happy with an existing visualization framework, Quasar comes with numerous dedicated connectors. It integrates perfectly with Microsoft Excel, Microsoft Power BI, Tableau, SAS Vaya, Grafana, and many more!
Integrating with an existing analytics toolset is the best way to ensure rapid adoption and a smooth transition.
Quasar is available on-premises, on the edge, and in the cloud. Learn more about our integrations in the partners page.
Quasar supports Intel x64 and ARM64 architecture for server deployments. Attached or detached storage architectures are both supported.
Docker and Kubernetes are fully supported.
Quasar supports Microsoft Azure, Amazon Web Service, and Google Cloud Platform.
For scenarios where you need a data infrastructure close to the source, you can deploy Quasar on embedded devices. ARM32, ARM64, and Intel-compatible architectures are fully supported.
