Platfora Shoots for a Big Data BI Layer
March 26, 2013 – Platfora has equipped its namesake in-memory platform for Hadoop to bring the interest in big data closer to business users and keep the complex data processes behind the curtains.
After two years of development and six months of beta testing, the Platfora release, launched today, is touted as a native, in-memory BI platform that rests on the application layer above Apache Hadoop, which operates as big data storage and processing operating system. The software platform works as a “reservoir” for business users, enabling visualizations and other analytics tied into Hadoop data, and carries capabilities for repeatable big data projects, says CEO Ben Werther. Platfora has also partnered with Hadoop distributions in Cloudera, Hortonworks, MapR, EMC’s Greenplum and others.
Platfora’s beta test included nearly two-dozen customers, like car sales specialist Edmunds.com and game maker Riot Games, though Platfora itself isn’t targeting any specific industries. The release is the culmination of the vendor’s efforts to expand access to data sets that can either be unattainable by most in the business, or come with a long lead time for connecting Hadoop to data warehousing or other enterprise data sources, Werther.
“There’s a generational shift going on in the way people are using data,” Werther said. “We’re seeing these groups of people who have adopted Hadoop and have this problem where they need a practical solution to let their business users get their work done.”
The platform launch also comes at the front of potential plans to expand Platfora’s operations and team beyond its current 50-member ranks, which was aided by a $20 million round of funding in the fall.
Two analysts briefed on the new platform see it as coming along at a time when the enterprise big data market is ripe. Mike Matchett, senior analyst and consultant with Taneja Group, says that the “secret weapon” in the Platfora release is the in-memory cache that automatically updates from the larger Hadoop data store, which “enables analyzing massive, varied and highly dynamic data sources.” This opens up both the speed of access to data in Hadoop clusters, but also handles back end data composition work away from business analysts and power users to avoid complication.
“Traditional BI solutions always required long lead times to develop ETL workflows and evolve data warehouses even over relational data. When looking at vast volumes and varieties of new big data, traditional BI approaches just can't keep up,” Matchett said.
Evan Quinn, senior principal analyst of data management and analytics for Enterprise Strategy Group, says the Platfora release addresses a few key medium and large business concerns with big data initiatives, such as the desire for a platform rather than a few software elements pulled together, as well as the deficit and high cost of expertise with Java MapReduce engineers or data scientists who are often needed to “make Hadoop work for BI and analytics.”
“A few others have tried, but Platfora has come up with an architecture that thus far uses Hadoop to best advantage, and uses Platfora’s own technologies to deal with aspects that Hadoop doesn’t natively do very well or at all, like data integration, complex query performance and visualization. In a sense, Platfora has taken excellent advantage of being a ‘second mover’ in the Hadoop-based BI/Analytics space,” he said.
Actually, Quinn says that the main challenge for the platform in this initial phase may be the vendor’s response to “overwhelming demand” from businesses primed to find big data returns.