IBM InfoSphere Streams

IMT_icon_Streams_84pxIBM® InfoSphere® Streams is an advanced analytics platform that allows a user to analyze data arriving from thousands of varied real-time sources with sub-millisecond response times.  Data is ingested by many operators (small modular elements), each of which performs a single specific analytical procedure that subsequently passes the processed data downstream to other operators.  Streaming applications are easy to develop using an Eclipse-based IDE and a functional programming language called SPL (Streams Processing Language).

InfoSphere Streams includes the following features:

Comprehensive Development Tools

  Graphical Editor: build applications by dragging & dropping operators while automatically synching graphical and source code views

Data Visualization: dynamically add new views to running applications with out of the box charts

Visual Application Monitoring: identify issues with an instance graph that displays application metrics


Scale-Out Architecture

Distributed runtime platform that scales from a single server to many nodes to process millions of events per second with microsecond latency

Big Data Integration

Hadoop Distributed File System (HDFS), and IBMs Watson Explorer

Sophisticated Analytics

  Geospatial: enable location based services with high performance analysis and processing of geospatial data

Time Series: generation (synthesizing or extracting), pre-processing (preparation and conditioning), analysis (statistics, correlations, decomposition and transformation), and modeling (prediction, regression and tracking)

R-Project Toolkit: data analysis with statistical, mining and modeling capabilities

Complex Event Processing: detect composite events in streams of simple events using patterns

SPSS: SPSS Analytics toolkit for Streams

Mining Toolkit: for the Predictive Model Markup Language (PMML) standard models


Text Analytics

Data flows through a stream in a tuple structure, which may comprise one or more text fields containing unstructured entries

Annotators written in AQL (Annotation Query Language) may be used to extract valuable concepts from the unstructured fields and create new structured fields in the tuple structure


Language Extensions

Native functions and new operators can be implemented to extend SPLExtensions can be written in SPL, Java or C/C++

InfoSphere Streams is shareware. The license restricts usage to development and testing only but contains most of the components found in the full InfoSphere Streams offering so users can explore the wide range of InfoSphere Streams functions.