Performance, scale, and real-time analytics: Red Hat JBoss Data Grid 7.1

24 mai 2017Syed Rasheed, Senior principal product marketing manager, Red Hat

I am excited to announce the general availability of Red Hat JBoss Data Grid 7.1! This is the only Red Hat software ranked highly in two separate Forrester waves categories: In-Memory Data Grid and In-Memory Database. On top of that, no other vendor offers any unified in-memory data management solution that is recognized in both waves -- JBoss Data Grid is the one product with the versatility to span both categories.

In-memory computing is all about high performance and scale-out architecture. The primary focus of this release to enhance the performance of JBoss Data Grid as an in-memory data management platform for hybrid transactional and analytical (HTAP) workloads.

New Capabilities and Features

Performance improvements. JBoss Data Grid 7.1 features core performance improvements, especially in clustered write operations. Current tests have shown up to 60% increase in write throughput under load. (We have modified various default settings to improve JBoss Data Grid performance.)
Elastic scale external state management for JBoss Web Server (Tomcat) and Spring applications (on-premise or cloud/Openshift). JBoss Data Grid 7.1 features the ability to externalize HTTP sessions from a JBoss Web Server node to a remote JBoss Data Grid cluster. This helps make the JBoss Web layer stateless and enables a rolling update of the application layer, while retrieving the session data from the JBoss Data Grid layer. Additionally, JBoss Data Grid 7.1 features Spring session support, which enables you to externalize HTTP session from a Spring (or Spring Boot) deployment to a remote JBoss Data Grid cluster.
Real-time analytics, through Apache Spark 2.x integration supporting RDD and DStream interfaces.
New string-based querying with Ickle (tech preview). JBoss Data Grid 7.1 introduces a new string-based querying language, Ickle, as technology preview, which enables you to specify combinations of relational and full-text predicates (based on Apache Lucene). This enhances the querying feature-set available in client-server mode by bringing several additional operations that were previously available only in library mode.
Ease of administration. Update and save node-level configurations are now available through the administration console.
Feature enhancements to Hot Rod clients, including streaming large-sized objects in chunks from the JBoss Data Grid server to a Java client and adding cross-site failover for C++, C# and Node.js clients.