Home > Why QuasarDB > What

What is quasardb?

Time series. At Scale.

Quasardb is a high-performance, distributed, column-oriented database with native time series support.

Transfers, computations and aggregations are so fast that analysts can work interactively on their data, regardless of the amount stored.

  • First class performance: never wait on the database with millions of requests per second per node
  • Direct to data approach for easy integration with the existing applications (Microsoft Excel, ActivePivot, Apache Spark...) or through our API
  • Transparent scaling to accommodate growth

A distributed key-value store

At the heart of quasardb is a unique distributed key-value store technology. It scales horizontally and vertically for virtually unlimited capabilities. All of this, without any compromise regarding reliability and consistency.

  • Peer-to-peer distribution based on the Chord algorithm
  • Lockfree and multithreaded code for vertical scaling
  • Implementation in modern C++ for maximum efficiency

Time series

You can store raw data of any form and size into quasardb, and you can work your on time series with native operations. Quasardb takes away all the hard problems such as concurrent access, reliable distribution, reduction, and efficient operations.

  • Time series, and data structures scale transparently and automatically across the cluster
  • All operations are atomic, durable and consistent by default
  • Billions of rows per second per core aggregation speed (using SSE4.2 and AVX2)

Transactions

How to manipulate several entries at once in a consistent fashion? Quasardb comes out of the box with distributed transactions. Transactions enable you to execute commands on several entries in a safe, predictable, consistent manner.

  • Distributed transactions based on Multi Version Concurrency Control (MVCC)
  • Unlimited read scalability
  • No global lock: transactions can work in parallel on different entries

Tagging and real-time search

How to find your entry when you don’t have its key? The typical solution is to create a schema and perform queries based on this schema. Quasardb has a better, schema-less approach: tags. Tags are extremely fast and scale automatically. Don’t compromise between search and speed: just tag!

  • Distributed tags for flexible data management
  • Immediate tag-based lookup
  • Tags can be removed or renamed

Fault tolerance

Quasardb is write-safe by default and offers features to sync absolutely everything you do for maximum safety. Thanks to its master-less design, quasardb is able to face the loss of nodes transparently. When a node joins back, quasardb will replay the log of events to guarantee consistency.

  • Fault-tolerant by design: perfectly symmetric peer-to-peer architecture
  • Safe by default
  • MVCC, high-resolution timestamps and the Chord algorithm are used together to guarantee consistency
Try now Contact