StreamBase

As its name implies, StreamBase is designed to handle the real-time processing of streamed data. Primary organizations targeted by the technology include financial services, telecommunications, the military, Homeland Security, and network monitoring and management applications.

The StreamBase Stream Processing Engine (SPE) is designed around the "inbound processing" concept. Whereas traditional data processing is typically centered around store-and-query methodologies, where data is first stored into a RDBMS and then queries are executed to extract data as it is needed (or continuously, immediately after is is stored); the StreamBase technology operates on the streamed data on-the-fly, before it is stored to the DB (if indeed it is stored at all). Additionally, the StreamBase SPE places each stream processing application in a single operating system process; to minimize inherent latency that can be the result of hand-offs between multiple data handling processes. For processes requiring the accumulation or aggregation of data before processing, StreamBase SPE includes integrated storage technology allowing the application to maintain state using in-memory hash tables, local disk, or remote databases via ODBC as needed. Integration of the stream engine to existing data feeds is facilitated through available interfaces including TIBCO Rendezvous, published Java and C++ APIs, and ODBC.

A cornerstone of the StreamBase technology is StreamSQL, an extension to standard SQL that provides three additional capabilities targeted specifically to streaming data processing. First, StreamSQL allows for the definition of user-defined "windows;" data queries that can be delimited by time, the number of messages, or other break-points. Second, StreamSQL provides the ability to deal with "Stream Disorder," or data that arrives at the application out of sequence. Finally, stream-specific operators are available for immediate use and can be extended via user-defined functions or aggregations.

Components of the StreamBase platform include the stream processing technology described above (the SPE), and an included development GUI with drag-n-drop design as well as testing and deployment capabilities (a built-in load simulator is available for testing designed apps).

New to the StreamBase offering is support for 64-bit processors; including AMD Opteron. According to the vendor, 64-bit support pushes the potential performance of the SPE to over 1 million messages per second.

Contact StreamBase Systems for further information.

product submission by DatabaseJournal Staff

Latest category updates via our RSS feed