Hadoop has a query problem. Map/reduce is dead. Spark is slow and complex. SQL has been idiosyncratic and slow.
Kognitio on Hadoop fixes that. Now you can use standard SQL to run huge numbers of concurrent in-memory queries over vast datasets. Suddenly the data in your Hadoop cluster is as easily queryable as your relational database.
Whether handwritten or tool-generated, Kognitio works with standard SQL queries as well as ODBC and JDBC drivers.
Even complex queries complete in real-time, making batch query a thing of the past. Kognitio is the ideal query engine for BI and data science.
In-flight query planning and dynamic machine code generation ensure every last CPU cycle is used in the most efficient way possible, enabling massively parallel processing of potentially thousands of complex queries.
Some engines fail when a query takes too much memory. Kognitio’s unique in-memory data structures and active resource allocation ensure that every query completes.
BI tools generate wide-ranging, complex SQL queries. Kognitio has been battle tested with common BI tools and by our own in-house team of data scientists.
Kognitio can run as a single server or a distributed cluster of thousands. Scale out as and when demand dictates.
Kognitio works as a drop-in component of your Hadoop infrastructure, allocating resource as a well behaved YARN application.
Kognitio on Hadoop is free-of-charge wherever and however you use it, with a growing community of developers.
Application & client layer
Kognitio analytical platform layer