What’s New in IBM InfoSphere Streams v3.1
IBM InfoSphere Streams V3.1 is a software platform that enables real-time analytic processing on Data-in-Motion for big data by providing comprehensive development tools, scale-out architecture and sophisticated analytics with toolkits and accelerators.
IBM InfoSphere Streams v3.1, released May 21, 2013, contains enhancements that include:
- Improved Java Performance - Several improvements have been made to significantly increase the performance of Java programs running on Streams including faster throughput and reduced resources for fused java operators.
- Performance improvements for operations using maps and - Compiler improvements eliminate the need to create a temporary copy of maps and lists for running programs, resulting in 2x to 10x performance increase for operators who perform functions on maps and lists
Development & Analytics Enhancements
- Support for R analytics - Streams now allows you to use R scoring in a native format while also taking advantage of the scalability and parallelism and partitioning of Streams to achieve exceptional throughput. R support extends the InfoSphere Streams family of analytic options.
- New IBM Time Series Toolkit operators - Existing VAR (Vector Auto-Regression) and ARIMA (Auto-regressive integrated moving average) operators support model retraining, monitoring, load, suspend, and resume operations. New operators and usability improvements enhance real-time and predictive analysis of regularly generated data from various sources including physiology, nature, systems or markets, and finance
- JUnit testing integration for Java - Java primitive operator developers can unit test their operators using JUnit or compatible test frameworks.
- Improved Streams Studio support - Developers can now use Java primitive operators within a Streams Processing Language (SPL) toolkit project including wizards and an integrated Java nature rather than a separate Java project.
- SPLDOC tool for the Streams Processing Language - This documentation tool provides simplified tools to generate documentation for Streams Processing Language toolkits and programs.
- Red Hat Package Manager support - Streams now provides an RPM spec file to create an RPM that can be used to replicate installation across many nodes, simplifying installation and maintenance.
- Separate data and control network support - Deployments can use a separate physical network for management services, allowing Streams administrators to control management services in the presence of high application data stream volumes that saturate the network.
- SUSE Linux Enterprise Server (SLES) - Streams now supports SLES on x86 hardware as an alternate platform for the runtime and development tools.
- Simplified instance recovery - By providing a single command for handling repair of an instance, Streams now checks for services that have failed and automatically restarts failed services in the proper order.
- Simplified restart of application components - Streams administrators can move or restart multiple application components on a new node, simplifying their work and speeding time to maintaining an optimal runtime.
- Support Zookeeper - Streams now includes the ability to use Zookeeper, an open source project that enables highly reliable distributed coordination, as a highly resilient name service, which improves availability.
- Streams Console visual enhancements - New IBM UI styles and widgets improves the look and feel of the administrative console.
- Java Message Service (JMS) support - The addition of the JMS interface extends the reach of Streams to communicate with more systems.
- Teradata and Asterdata support - Streams supports exchange of data between Teradata and Asterdata databases and warehouses.
- Representational State Transfer application programming interface (REST API) - This public API for management and monitoring of the system allows callers to get state and metric information for the system and applications.