Vertica, a column-oriented analytic database technology provider, has announced a new release of its Analytic Database product with new features for performance gains, increases in scale, and advanced analytics. The primary new features consist of a next-generation column-store architecture called FlexStore, MapReduce technology for advanced ETL loading and analytical processing, and load balancing via IPVS.
"The FlexStore technology within version 3.5 of the Analytic Database improves performance because it reduces the number of reads that must be done in various queries," Dave Menninger, vice president of marketing and product management at Vertica, tells 5 Minute Briefing. "This is accomplished by grouping columns of related data that are always accessed together next to each other, and having queries that require this data read it as though it were located in just one column (i.e., one read versus two). FlexStore also consists of intelligent disk utilization technology that automatically ranks disk storage devices based on I/O rate performance and puts data that is accessed most often onto the fastest disks."
Version 3.5 also includes Map Reduce technology, which is an analytic technique to exploit parallel processing. This technique splits analytical activity across multiple compute nodes (i.e., mapping different computations to different computers) and then reducing all of the calculations down to the final result set. Two versions of Map Reduce are supported-the Amazon Elastic Map Reduce (EMR), and the open source technology named Hadoop.
Other features include the ability to utilize commodity grid hardware, leverage existing ETL and reporting tools, a new data type to support certain numeric fields for Microsoft and Oracle databases, and support for Perl and Python scripting languages. With this additional scripting language support, developers can do things such as create batch jobs to pull data from web logs and load it into the vertica Analytic Database.
For more information on Vertica Analytic Database 3.5, go here.