At the Strata Conference today Calpont anounced InfiniDB 3, the latest release of its high performance analytic database. Designed from the ground up for large-scale, high-performance dimensional analytics, predictive analytics, and ad hoc business intelligence, the new release includes capabilities to capitalize on a variety of data structures and deployment variations to meet organizations' need for a flexible and scalable big data architecture, according to the company.
A key enhancement in the new release is simplified big data loading with new parallel data load options, Nick Ochoa, vice president of marketing, Calpont, tells 5 Minute Briefing. "We already had a very fast data loader but now we are going to improve that in an expanded MPP model - and that is really important for many of our customers, especially those in the telecommunications and content delivery networks."
When customers are dealing with massive data sets, time is of the essence on the load side, once the query performance side has been addressed. "We are now seeing loading of up to one terabyte of data per hour," Ochoa says. "Calpont was already on the bleeding edge and we are just enhancing that based on the requests of customers that needed even more bandwidth for data load capacity."
With the new release, data load options range from a simple single bulk load command to parallel bulk load commands that treat the data source as independent or shared. The automatic data distribution can occur across a suite of shared-nothing disk resources and grow linearly with tiers of data sources, files and commands.
In the second key aspect of the new release, Ochoa says, InfiniDB 3 provides provisioning and run-time operations on Amazon Elastic Compute Clouds (Amazon EC2). Prepackaged Amazon Machine Images (AMI) help speed deployments with the automatic provisioning of InfiniDB nodes on EC2. For organizations that expect rapid growth, the transparent support is built in for EC2 virtual storage and data redundancy to meet Amazon Elastic Block Store (EBS) polices. "There is definitely a push to [move to the cloud] because the cost and maintenance of hardware on-premise is becoming more of a challenge," Ochoa observes.
"We have customers today that have deployed InfiniDB in an on-premise capacity and now are looking to the cloud, so providing an Amazon Machine Image for InfiniDB enables our customers to get flexibility out of what they are already benefiting from today," he says. "And for those who are not customers yet, and are now looking for cloud deployment this is a starting place for them so they can actually go and do testing within an Amazon environment and then if they decide they want to go move forward with the product, they can just deploy it based on the testing they have already done."
InfiniDB is a massive parallel processing (MPP) columnar solutionthat executes SQL written queries as parallelized map and reduction operations. When companies query billions of data points, InfiniDB's two-tier distributed architecture optimizes query processing to all available CPU threads, minimizing I/O reads from disk, and accessing only the data required for analysis. According to Calpont, InfiniDB is ready for large-scale workloads out-of-the-box, requiring no indexes, materialized views, manual tuning, or manual partitioning.InfiniDB 3 is scheduled for general availability in April. For more information about InfiniDB or to try the Enterprise version, visit www.calpont.com.