MonetDB Solutions has announced the release of MonetDB/e, a high-performance embedded version of the MonetDB database system for data analytics. It brings the full power of the world-renowned column store MonetDB for business intelligence, data analytics and machine learning into the heart of your application. MonetDB/e combines the best from both embedded and server-based RDBMSs into one system. It delivers what data scientists have been waiting for: a lightweight and easy-to-use embedded database offering fast analytical queries and support for transactions. MonetDB/e is backed by world-class support and database expertise from the engineers who built it.
“Who would have thought that you can package a versatile and powerful RDBMS for data analytics into something you can simply implant into your application code and boost performance,” said Martin Kersten, CEO of MonetDB Solutions and world-renowned database researcher. “We managed to reach the point that the users need not think upfront about complex client-server infrastructures.”
Traditionally, embedded database systems focus on offering a small footprint and convenient processing of a few data records. When faced with larger data sets, developers are quickly forced to switch over to a server-based solution. This requires considerable management and maintenance effort. MonetDB/e offers what other embedded RDBMSs cannot, packing a highly optimised query processing engine into a tight core which can then be easily embedded in various programming environments. MonetDB/e delivers:
High-performance embedded data analytics: MonetDB/e inherits performance from MonetDB, well-known for its lightning-fast speed in processing analytical queries. Its highly parallelised execution engine thoroughly exploits modern multi-core platforms.
Zero-cost data exchange: the columnar storage model of MonetDB/e has the same binary structure as, e.g. NumPy arrays and R DataFrames. Data exchange with the language environment is as simple as flipping a pointer.
In-memory or persistent: MonetDB/e supports both an in-memory-only mode for higher speed and an on-disk mode for data persistence. Developers switch seamlessly between these options by changing just the connection string.
Painless deployment: MonetDB/e is the only embedded SQL engine with a code-compatible accompanying server-based version (i.e. MonetDB). By changing the connection string, developers can switch from MonetDB/e local store to a (remote) MonetDB server. This feature minimises the effort needed to migrate from an embedded database for development to a multi-server database system.
Easy migration: MonetDB/e supports an interface fully compatible with SQLite3 and Python DB API-2. Developers can choose the most convenient system by just changing the name of the driver library.
Support from the source: MonetDB/e is backed by the developers of MonetDB and their decades of experience in database engineering, support and research.
Customers deploying MonetDB in their business infrastructure represent a wide range of markets, including infrastructure management, bio-genetics, network monitoring, search-engines and the automotive industry.
“We partner with some of the largest organizations in the world to deliver digital workflows that create great experiences and unlock productivity,” said Pat Casey, Chief Technology Officer and Senior Vice President of DevOps at ServiceNow. “Our priority is to make the Now Platform the most innovative for our customers to get work done. We invested in MonetDB to share our expertise in cloud automation and drive greater benefits for some of the world’s largest companies.”
“MonetDB/e is a great addition to our software development toolbox,” says Arjen de Rijke, CTO at DataSpex. “Our complex data processing pipeline is already running on MonetDB servers. With MonetDB/e, we have integrated unit tests for all queries in our CI environment. This largely reduces our development effort and shortens the deployment circle.”
“The tailored search technology developed at Spinque has been empowered by MonetDB innovations for over a decade,” says Roberto Cornacchia, CTO at Spinque. “The MonetDB/e release marks an exciting new direction for even tighter and more efficient integration.”
“What these users have in common is the need for fast and reliable data analytics at scale,” says Ying Zhang, COO at MonetDB Solutions. “It’s why this kind of database technology is ideal for interactive data analytics and the new ‘playing fields’ it opens up.”
MonetDB/e is part of the open-source MonetDB software suite and is distributed under the liberal MPL2.0 license.
MonetDB/e is available on all major operating systems including Linux, Mac and Windows. The current release supports the Python and C language bindings. Support for R and Java is expected in early 2021. The Python package monetdbe is available from PyPi. Users won’t need to install or compile the MonetDB code itself.
Users interested in the technical features can see an example of using MonetDB/e to analyse New York taxi trips here.