Published On: Wed, Jan 6th, 2010

The Merits of Kx Systems

KX SystemsKx Systems supplies many of the world’s major banks, hedge funds, exchanges and other financial institutions with a unique and highly original solution set for ultra-high performance database and time-series analysis.

Financial institutions are facing constant pressure to collect, store and analyse data from numerous sources, such as exchanges, in-house or vendor. Not only are data volumes increasing at near exponential rates but institutions are facing tighter regulation, coupled with a drive for greater transparency. How can transparency and risk management be achieved without knowing the current position, which in turn can only be done with sufficient processing power?

Kx products are designed to meet the requirements for efficiency, transparency and risk management. Its products are used to analyse vast quantities of data in real-time: Kx offers the kdb+ database and q, its proprietary high-level language, which enable institutions to process constantly increasing data volumes. Kx’s products are being continually optimised to take full advantage of the latest technology, such as Intel’s Xeon processor 5500 series, ensuring high performance and reducing latency to a minimum. With a single data format for streaming, real-time and historical data, kdb+ provides an end-to-end database that eliminates latency across multiple data management systems.

Since its inception in 1993 Kx has led the market with new functionality and features. For example, Kx was the first to introduce nanosecond timestamps in 2009, which allow for easy identification of the most recent events, such as trades, enabling much faster and simpler processing and letting users perform more complex operations, creating a competitive advantage.

Some of the benefits offered by Kx are:

• Production applications can be delivered in as little as 24 hours from installation, providing a fast return on investment (ROI).
• Queries against streaming, real-time and historical data, or a combination of any of these. The ability to treat streaming and disk-based data as a single database for analysis ensures high processing speeds and supports a variety of trading, risk management and compliance/regulatory applications.
• Fast data capture and analysis even as data volumes continue to rise, with a scalable path for growth.
• Embedded multi-threading capabilities allow developers to take advantage of parallel processing without changing their code.
• Takes advantage of multi-core architecture in order to handle the necessary volume, volatility and speed of data updates and analysis.
• Compatibility with legacy technologies.

Kx solutions enable trading groups to implement sophisticated trading strategies in real-time, no matter how much the volumes of data increase or fluctuate, with the ability to capture more than one million messages per second. In handling such a large number of updates, Kx also delivers the headroom to go far beyond simply capturing and updating market data. For example, it can simultaneously support many thousands of real-time custom queries, proprietary analyses on data in-memory, update database indices dynamically to speed query execution, and log data to the file system. Once saved to disk, massive amounts of data can be queried at speeds of ten million records or more per second (per core). Saving streaming data allows institutions to test their strategies and mine market data history for the trading intelligence that creates a competitive advantage.

About the Author


Displaying 1 Comments
Have Your Say
  1. Steve says:

    From the High Frequency Group on Linkedin:

    Although Kx is cool I fail to see why people pay for this. It seems to me it is a nice-to-have as opposed to a must-have product. The cost is high as you have to train someone in the query language and kind of have to have a dedicated programmer to run the thing.
    By Vincent Mayeski High Frequency Quant Trader and Technologist

    As noted in the article, Kx is geared towards high-performance, high-volume databases, and companies pay for it because it gives them a competitive advantage. The query language used is essentially a superset of SQL and is easily learned by anyone with a background in databases. Dedicated programmer support is not needed for running the database, though typically companies do implement analytics directly in the query language, so a knowledge of the language is useful.
    By Steve Edge Electronic Trading Disciple

    Companies do pay for Kx because it gives them a competitive advantage, in part because of its performance in processing high volumes of data, but also because of its ability to implement analytics, such as a trading strategy, with remarkable speed. However, it is not “essentially” a superset of SQL and does require a dedicated programmer(s) if you want to utilize its full power. Q, the query language, is a variant of APL with built-in support for time intelligence and handling tabular data. The language can mimic basic SQL syntax and someone with a database background could use it to perform basic data access, but this is only a fraction of what Q is capable of and the latter does require an expert programmer. This is not a criticism, Kx is a great product and there is nothing else in the market that can do what it does. The competitive advantage is real, which is why you see it used by so many hedge funds. But you’re fooling yourself if you think this doesn’t require a commitment of resources.

    I know that one thing that’s easy with it is loading up of TAQ data. It does queries on TAQ very fast as well. So if you are looking for a place to store TAQ and you have a big budget this is probably the way to go. I cant see why one would want to deploy a trading strategy on top of it however.
    By Vincent Mayeski High Frequency Quant Trader and Technologist
    By A B Carter Senior Developer at Canada Pension Plan Investment Board

Leave a comment

XHTML: You can use these html tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

IRP Journal

IRP Journal



Poll results are published in our Weekly Newsletter -->
All Rights Reserved WIld Wild Web Limited