Sybase RAP is a unified real-time analytics platform that lets capital market firms make better trading and portfolio decisions across the trade lifecycle — from better model development to execution with less risk through timelier, more comprehensive market insight. Sybase RAP combines the power to manage massive data volumes with the speed needed for real-time data analytics.
Sybase RAP combines real-time tick capture, trade capture, and historical analytics in a single integrated platform. It includes high performance data stream capture, an in-memory transactional cache for immediate availability of new information and fast query speeds, and a highly scalable historical store that is optimized for analytics.
What's new !
Sybase RAP version 4.1, provides support for R.
RAPStore lets users invoke R programs running on an R server as if they were SQL functions, using the RAPStore UDF infrastructure. In SQL databases, a user-defined function (UDF) is a mechanism for extending the functionality of the database with custom analytics. RAPStore allows UDFs written in C/C++ to be invoked like any other SQL function. Those UDFs can in turn invoke R programs. This broadens the analytical capabilities of RAPStore in an open, extensible way. RAP users benefit from the rich set of existing statistical packages available within the R community, and can make use of them for application development.
R users can query data in RAPStore using RJDBC. This can be used to push data pre-processing activities into RAPStore where it can often be accomplished much more efficiently. Pre-processing operations, such as data cleansing, filling in missing values, and so on can be written as stored procedure calls within RAPStore. This enables only the relevant results to be downloaded to the analyst workstation, as opposed to downloading all the data from the database. Additionally, if there are many R users that need to access data for their analysis, common preprocessing functions can be shared in RAPStore, ensuring that data cleansing and pre-processing operations are performed consistently for all users. Finally, RAPStore allows many terabytes of data to be stored efficiently with significant compression (50-70%), providing R users with large amounts of data to effectively backtest their algorithms.