Big Data management software specialist RainStor Wednesday debuted its new native Hadoop database, RainStor Big Data Analytics on Hadoop.
The enterprise database is intended to give organizations the ability to run fast and flexible analytics on multi-structured data without having to first move that data out of the Hadoop Distributed File System (HDFS) environment.
"To manage growing, complex volumes of data requires more efficient data management to drive down infrastructure costs and at the same time provide flexible, high-performance analysis," said John Bantleman, chief executive officer of RainStor. "RainStor uniquely delivers the flexibility to use both MapReduce and rapid-response SQL query, which gives customers the ability to get more out of their Hadoop environments. RainStor on Hadoop essentially gives the enterprise high-performance analytics at sale, without the premium cost of a data warehouse."
It's important to note that RainStor's database is not an operational database.
"We're all about enterprise Big Data management," said Deirdre Mahon, vice president of marketing at RainStor. "Don't think of it as an operational database or a pure data warehouse. It's designed to handle large volumes of raw data."
RainStor may not be a familiar name to many, but it has been in business since 2004 and selling commercially since 2007. However, until last year, it sold only to original equipment manufacturers (OEMs) that embedded its database technology in their products. In 2011, it decided to start a direct business in addition to selling to OEMs.
"We were getting demand from large customers, specifically telecommunications, which is a large market for us," Mahon said. "They wanted to buy from us directly."
She explained that RainStor has about 150 deployments around the world, typically with large, blue-chip companies. It's largest installation today is a Japanese telco that ingests up to 17 billion records per day.
According Mahon, what sets RainStor Big Data Analytics on Hadoop apart is compression. She said RainStor achieves a reduction of up to 40x (97.5 percent) or more compared with raw data, and that data requires no re-inflation when accessed.
"We're usually 10x better than even columnar databases," Mahon said. "Because of the compression, you don't need as many nodes as your data volumes grow, so you can get a major cost reduction."
She explained that RainStor is capable of ingesting data off networks at a rate of tens of billions of records a day, and it compresses the data at the same time it ingests it. Then, by combining this compression with dynamic filtering at file, column and row level, she said RainStor achieves faster analytics from more efficient use of the Hadoop cluster. In turn, she said, this gives the ability to run faster query and analysis using both SQL query and MapReduce with 10-100x faster results.
"A lot of our customers, particularly in the telecommunications sector, have tens of billions of records coming in off the network every day," Mahon said. "A relational database can't ingest it that quickly. A high-end data warehouse quickly becomes prohibitively expensive."
By combining compression with RainStor's enterprise data management features, the database also reduces storage and cluster size for lower operating costs. RainStor said the reduction of nodes in the Hadoop cluster can lead to about 85 percent lower operating costs.
"As data volumes continue to grow and users look to take advantage of both structured and unstructured data, offerings like RainStor Big Data Analytics on Hadoop that enable co-existence without data movement are likely to come into greater focus," said Matt Aslett, research manager of Data Management and Analytics at research firm 451 Research. "RainStor's de-duplication and compression capabilities also offer the potential to lower cluster size and cost, while improving analytic performance, which will be key considerations as we see even greater adoption of Hadoop."