Oracle has released the first version of the Coherence in-memory data grid software since it bought the vendor of that software, Tangosol, earlier this year. Over time, Oracle hopes to dramatically widen the appeal of the product.
Data grid software is middleware that manages data objects in-memory across multiple servers. By storing data in memory and not in slower back-end systems, it is able to provide users with very fast access to frequently used information. It helps enable what’s been dubbed extreme transaction processing (XTP), the use of highly scalable, secure and reliable applications which generate vast amounts of transactions to give access to real-time data.
Oracle Coherence 3.3 officially debuted earlier this month after Oracle incorporated feedback from early users who’d been working with the software over the previous two months, says Cameron Purdy, vice president of development in Oracle’s Fusion middleware group. The new release features improved performance and clustering, tighter links into Fusion, and support for Microsoft’s .Net framework.
Coherence competes against IBM’s ObjectGrid software and products from GemStone and GigaSpaces, Purdy says. Oracle bought Tangosol in March and Tangosol’s staff joined Oracle. Purdy is the former CEO of Tangosol.
While Coherence has proved very popular in the financial services sector, the middleware and the XTP concept is still viewed as somewhat exotic among other industries, Purdy says. “We want to take this exotic technology and put it in the hands of every developer,” he says. “If we do our job right, terms like ‘grid’ and ‘XTP’ will go away and [instead] will become the default technology.”
As with many of its purchases, Oracle is operating a two-track development roadmap for Coherence.
On the one hand, it will integrate the data grid technology with more and more Fusion components and with its TimesTen in-memory database and enterprise-level Oracle Database. The plan is to end up with a middleware stack that can support applications carrying out real-time data analytics, grid-based in-memory computation and high-performance transactions. In Coherence 3.3, Oracle has begun the process of tight integration with Fusion elements, including hooks into Oracle Application Server and Oracle TopLink, an object-relational mapping tool for Java developers.
On the other hand, Oracle has committed to maintaining the stand-alone version of Coherence, which isn’t hooked into Fusion. The current Coherence customer base runs a wide variety of different vendors’ middleware, with no product particularly dominant, Purdy says. The supported middleware includes Apache Tomcat, BEA’s WebLogic, IBM’s WebSphere, Microsoft’s .Net and Red Hat’s JBoss. Coherence 3.3 has expanded its support to Java Standard Edition (SE) version 6 and includes a revamp of the software’s Coherence for .Net offering.
In the future, as well as continuing to support connections to Java and .Net middleware, Oracle is keen to meet user demands to open up Coherence still further. Financial services customers are keen for Coherence to also be able to manage their C and C++ applications, Purdy says, and for the software to handle complex calculations created in a variety of different programming languages.
Coherence 3.3 comes in three versions — a standard, enterprise and grid edition.