GemStone Systems has been in the business of middle-tier caching since at least the early Java days. Like other players [for example, Azul, GigaSpaces, Tangosol] that have found success selling into low-latency financial and related systems markets, GemStone is looking to support mixed C# .NET and Java development.

For some of these firms, the interoperability journey includes C++ as well. In GemStone’s case, heterogeneous language support now means Java, C++ and .NET. Earlier this year, the company added native support for C++ and .NET clients to its GemFire Enterprise 5.0.1 “Enterprise Data Fabric.” The company said this eliminates the need to deploy a Java Virtual Machine (JVM) on an application client or to use wrappers when sharing data between C#, C++ and Java applications.

GemStone Systems has been in the business of middle-tier caching since at least the early Java days. Like other players [for example, Azul, GigaSpaces, Tangosol] that have found success selling into low-latency financial and related systems markets, GemStone is looking to support mixed C# .NET and Java development.

For some of these firms, the interoperability journey includes C++ as well. In GemStone’s case, heterogeneous language support now means Java, C++ and .NET. Earlier this year, the company added native support for C++ and .NET clients to its GemFire Enterprise 5.0.1 “Enterprise Data Fabric.” The company said this eliminates the need to deploy a Java Virtual Machine (JVM) on an application client or to use wrappers when sharing data between C#, C++ and Java applications.

One of the reasons to forgo the JVM, no doubt, is to meet the higher performance and scalability requirements some companies now have.

GemStone has been in these types of business for some years, and originally fashioned its technology around the notion of the Object-Oriented Data Base[OODB]. We talked recently with Jags Ramnarayan, chief architect. GemStone.

“In the 1990s, we did an early J2EE application server,” Ramnarayan said. “We took our object database and data binding technology and did it as a middle-tier caching solution.”

After reviewing various customer engagements in some high-performance settings, GemStone began to focus on eliminating data sharing latencies between processes – by that time it was a mix of Java and C processes.

This occurred after the advent of the Jcache standard, one that proved somewhat lacking. “Jcache as a standard never went anywhere,” said Ramnarayan. “Users had to build a lot of the functionality on their own.” The term that GemStone settled on to describe what was needed was the “Enterprise Data Fabric.”

This fabric, said Ramnarayan offers distributed caching and the ability to manage objects using a highly concurrent [in-memory] structure. Furthermore, it selectively offers RDBM semantics. [“Things like the ability to do asset transactions, synchronous or asynchronous persistence through ‘shared-nothing’ architecture, and the ability to do querying.”] As well, the fabric is event driven, comprising publish-and-subscribe notifications. As such, it may be in the vanguard of an increasingly popular middleware architecture. Finally, continuous querying – after all, the data in the systems GemStone is focused on is continually changing - is part of this ‘fabric’ approach.

Read the rest of the story on TheServerSide.NET

This entry was posted on Tuesday, May 15th, 2007 at 11:57 am and is filed under Interoperability, Distributed objects. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.