InfiniDB for In-Database Analytics

There are an emerging number of big data problems that can best be solved with a scale-out approach. For the big data problems that require custom code to solve vendor specific issues, then a toolkit for MapReduce may be the answer. For deep analytics of the data to determine behavior patterns or for session analysis, InfiniDB Enterprise for In-Database Analytics is a software platform solution that is easier to use and with significantly more out-of-box functionality.

Introduction to Map Reduction Operations

InfiniDB is a column database has an integrated map reduction implementation that enables a similar distribution of work to MapReduce in support of query processing. InfiniDB’s map reduction implementation transforms each query into a small number of map and reduction jobs that are executed automatically. The distributed nodes executing the reduction step all operate against dynamically assigned partitions of data and without any node-to-node communication. This distribution of work model is a key component to InfiniDB.

InfiniDB’s distinction is the tight integration of SQL functionality with the map reduction implementation. It offer filters, expressions, inner joins, outer joins, multi-table joins, correlated and non-correlated sub-query, group by, and aggregation behaviors as fully distributed reduction operations within its framework. This allows for complex queries (up to 60 table joins) to be requested via SQL and executed with the integrated map reduction process, rather than having to code new Erlang functions.

Distributed UDFs leverage integrated Map Reduction

UDFs are fully parallel and distributed, and run as an integrated operation within the InfiniDB database storage engine. This enables implementation of custom functionality that leverages the full benefits of InfiniDB's integrated map reduction capability, distributing the custom functionality across all available cores within the distributed layer. Further, InfiniDB provides user defined functions (UDFs) that allow for custom C++ code to be implemented as a fully integrated operation within the reduction framework. UDFs are fully parallel and distributed, and run as an integrated operation within the InfiniDB storage engine. This enables implementation of custom functionality that leverages the full benefits of InfiniDB's integrated map reduction capability, distributing the custom functionality across all available cores within the distributed layer.

Experience for yourself why InfiniDB’s fully integrated map reduction operations and UDF’s provide the best solution for in-database analytics – Try InfiniDB today.