GigaSpaces Targets Clouds for Real-Time Analytics of Big Data
GigaSpaces this week rolled out a new version of its application virtualization software designed to process massive amounts of streamed data and allow users to analyze it in real time.
The company's XAP 9.0 can run in an enterprise datacenter or within a public cloud environment. The software is designed to take the proliferation of big data coming from various sources -- such as social media and financial and Web transactions -- and allow business users to extract and analyze that data in real time. Founded 12 years ago, the Israeli-based company with U.S. headquarters in New York is well regarded for its in-memory processing technology.
Nati Shalom, GigaSpaces' founder and CTO, described XAP 9.0 as an in-memory storage and processing engine designed to process large volumes of data much like Facebook and Twitter do for their own analytics services, though those social networking giants have built their own custom real-time event processing platforms. Shalom said most organizes don't have the resources and skills to build such infrastructure, yet many have found themselves needing to contend with the huge amounts of data that have become commonplace to analyze in the past two years.
"We solved the problem of processing large volumes of data by processing -- I mean, from the time you get the feeds to the point you get the content generated," Shalom said in a telephone interview.
The new XAP 9.0 software is designed to process large amounts of data coming from various data types based upon various formats and languages. It is also designed to "keep up with the speed of incoming data, while maintaining real-time latency levels," the company said.
It does so using its real-time streaming parallel processing engine, which the company said uses "fine-grained" data compression, a reduced memory footprint that doesn't affect query capabilities and support for any cloud computing environment including Amazon Web Services, Microsoft's Windows Azure platform, services based on emerging standards such as Cloud Foundry and OpenStack, and services (and private clouds) from just about any provider, Shalom claimed. It also supports bare-metal environments for extreme I/O, he said. Though it's Java-based, XAP 9.0 can process code developed in C++ and Microsoft's .NET and other languages, Shalom noted.
Among the various big data sources it can process are Cassandra, MongoDB, SQLR and others. XAP 9.0 will initially support the Hadoop distribution from IBM called InfoSphere BigInsights with support for the popular Cloudera and Hortonworks Hadoop distributions to follow, Shalom said.
The software starts at "a couple" of thousands of dollars per core, Shalom said.
Posted by Jeffrey Schwartz on 05/24/2012 at 11:32 AM