Oracle is polishing off its Coherence data grid middleware with features aimed at enabling large-scale data analytics across distributed systems.
Oracle Coherence 3.6, a component of Oracle Fusion Middleware 11g, adds session management for ASP.NET applications, increased security, the ability to update live data in memory, and expanded support for transactional data.
Oracle acquired its Coherence technology from Tangosol in 2007. Coherence enables distributed applications to perform grid-based in-memory computations, real-time analytics, and scale out for high-performance transactions.
Coherence 3.6, released Monday, brings its .NET support up to speed with its Java counterpart with session caching. “In the Java space, Coherence has had session management for high-scale systems, but if you built a .NET application, you were stuck,” said Cameron Purdy, vice president of development at Oracle.
.NET developers have had to store session data in SQL Server or to use Microsoft’s ASP.NET StateServer, which doesn’t scale and has a single point of failure, Purdy explained. “Now [with Coherence’s Web-session management module], .NET servers can work together in a peer-to-peer manner that scales and eliminates that single point of failure.”
Oracle has added a feature called Coherence Quorum, which works across platforms to give applications more control over grid resources. An application may now override the default behavior of a grid, so that customers may address specific concerns in high-end applications, Purdy said.
All data that passes through Coherence is now secured by end-to-end Secure Sockets Layer connectivity. Coherence also provides user-level security, giving each user a private view of data that is being managed by the data grid, Purdy added.
Oracle also introduced Coherence Query Language (CohQL), an SQL-like language for developers that allows data to be selected and updated in a running system. Prior to this release, developers had to wait for information to be flushed out to a database, Purdy said. Developers can use CohQL to manage and use Oracle Coherence rather than using programming languages.
Further, Coherence now works across distributed transaction environments using Multiversion Concurrency Control. An XA adapter, based on the Open Group’s XA specification for distributed transaction processing, integrates data from IBM WebSphere MQ, Oracle and TIBCO systems.
Future releases of Coherence will focus on providing more predictability in managing large data sets in memory, and will leverage solid-state drives and flash drives as a form of “cheap memory” for backing up data, Purdy said. Those changes would provide for greater utilization of hardware resources, he added.
