Teradata, the data warehousing technology vendor, has announced its intention to acquire Aster Data, a company that applies ‘massively parallel processing’ (MPP) to conventional databases.
Aster Data’s technology is built on MapReduce, an MPP system developed by Google that splits large analytical workloads into smaller tasks and runs them in parallel. The company’s products combine MapReduce with SQL databases, in which most conventional business data is stored.
Teradata bought an 11% stake in Aster Data in September 2010, and today announced its intention to acquire the remaining stake for $263 million.
MPP is an increasingly popular technology for analysing particularly large data sets. It is used in applications such as web traffic analysis or fraud detection where, rather than taking a sample of the data, the analysis must include billions of individual data points.
In August 2010, Information Age spoke to Martin Willcox, Teradata’s director of platform marketing for Europe about MapReduce. He said it was “highly complementary to what we do”, but that “our experience is that [analysis] is not the optimal use for it”.
Willcox argued that MapReduce was best used for processing large volumes of data, but that once processed it is better to analyse that data in a SQL database. Integrating the two database technologies is precisely what Aster Data’s products do.
The analytics market has seen a number of acquisitions in the last year, such as EMC’s acquisition of Greenplum and IBM’s $1.7 billion purchase of Netezza. Last month, Hewlett-Packard effectively reset its business intelligence strategy, discontinuing its failed NeoView data warehouse range and acquiring Vertica, another so-called “big data” analytics supplier.