McObject Boosts Time-Series Data Chops

McObject, an Issaquah, Wash.-based provider of embedded in-memory databases, is preparing to roll out an updated edition of its eXtremeDB embedded database tailored for financial markets with new features designed to enable more efficient processing of time-series data, in response to growing data volumes and continuing demand for low latency.

The eXtremeDB Financial Edition is currently in beta testing, and is scheduled for general availability in the third quarter of this year, says Chris Mureen, chief operating officer at McObject.

Financial Edition is based on the vendor’s core eXtremeDB In-Memory Database System, and consolidates features from other editions within the eXtremeDB product family—including its Cluster, High Availability, Transaction Logging and Embedded 64-bit Database editions—as well as new features specifically developed to meet market data needs, says Ted Kenney, director of marketing at McObject.

A key new feature is support for a column-based data layout, which is typically how time-series data—such as tick data—is stored, in order to enable faster storage and retrieval of high volumes of similarly formatted data items than when stored in rows, Kenney says. In addition, the Financial Edition includes a library of vector-based statistical functions—including co-variance, weighted sum, minimum/maximum and standard deviation—and the ability to add user-defined functions that can be executed over one or more columns to support financial calculations for quantitative analysis.

In particular, Mureen says many beta users of the Financial Edition have deployed it to support back-testing servers, which capture high volumes of data from multiple exchanges that must be processed quickly.

The new edition also includes the graphical user interface and API for monitoring activity within the database, which McObject released earlier this quarter (IMD, April 23), as well as the vendor’s clustering technology, which distributes processing over multiple hardware nodes. The Financial Edition also supports the ability to define local tables for individual nodes, so that a task can be assigned to a specific node in a larger cluster, and the data to perform that task does not need to be duplicated across all the nodes.

With the Financial Edition’s support for time-series data, the vendor can extend its offering at its existing financial clients—including exchanges, trading firms and technology providers, which have deployed eXtremeDB to leverage its transactional capabilities and performance—as well as approach prospective clients that require time-series support, Mureen says.