Home > Why QuasarDB > Philosophy

Relational databases

For more than 30 years, relational databases enjoyed an absolute reign. If for some reason, you needed a database, the answer to that question was “you should use a relational database”. If your application is using a database, it is very likely that it is relational. Relational databases while extremely flexible and capable, have intrinsic limits:

  • Before storing a single byte, they need to be told how the data should be organized
  • Scaling is expensive and complex
  • Simple operations do not deliver the throughput modern hardware can

If you had to design a database…

…from the ground up, how would you do it? What would you do? At quasardb we started from a clean state and asked ourselves how we could push the envelope by being:

  • Simple and pleasant
  • Fast and frugal
  • Safe and secure

NoSQL

NoSQL databases emerged as the status quo for relational databases began to be questioned. Many have been let down by new cutting edge products and many technologies are actually running after relational features and forgetting where they come from. Quasardb is a key-value store with advanced features and actually delivers on its promises.

  • Truly-schemaless design for rapid development, outstanding performance and scalability
  • Reliable and simple mechanisms
  • Advanced features such as distributed containers, tags and transactions

Progress, not regression

If you remove all safeties from a database and forget about consistency or reliability, yes, it is very easy to be faster and simpler than a traditional relational database, but is it what you want? When we designed quasardb, we made sure that we would offer more, not less.

  • quasardb is safe, consistent and transactional like any database should be
  • Speed and scalability come from design choices, not shortcuts
  • Implementations details are abstracted: just store and access your data!

Big Data

Your data may not be big yet, but it may grow. With its scalable architecture, quasardb makes sure you are not losing opportunities by throwing away data. Because it is truly masterless, quasardb is also perfectly suited to store data en masse and work on it later. Quasardb will be ready when you are.

  • Virtually infinite scalability to suit any Big Data need
  • Actual, real-time lookup with tags
  • Hadoop integration for advanced analytics

CAP

One of the most frequent question we get is about the CAP theorem. Here is our answer: you no longer have to care about the CAP theorem. From the user point of view, the database is always consistent. If a partition occurs, we keep the data consistent and solve conflicts for you.

  • Last-write wins approach
  • Multiversion, high-resolution timestamp based conflict resolution
  • The engine decides automatically between consistency and availability to offer an uniform view

Open source

quasardb uses open source and we’ve shared all improvements we made on the library we use. Many of our API are also open source with a BSD license. The heart of our technology however is not open sourced as we felt it would make us lose focus which would ultimately hurt you, our users. If time permits we will open source more. Find us on github!

  • BSD-license APIs
  • Improved open-source elements shared back on GitHub
  • Open source initiative to package the most useful components

A cooperative approach

The temptation might be big to say that you should replace your RDBMS with NoSQL databases. In some cases, that may be the thing to do. In some other, that can end up being extremely costly and inefficient. We’ve made sure that quasardb can be used collaboratively with existing systems.

  • Use your data “as is”, we don’t impose any format
  • A wide range of open source APIs are provided
  • Data is stored in an open format with everything provided for offline extraction
Try now Contact