Hybrid transactional/analytical processing (HTAP) leverages the power of in-memory computing to bring OLTP and OLAP onto one platform. Before in-memory computing, transactional and analytical processes were split onto two separate platforms to reduce load on the transactional database. OLTP was run on an operational database. The transactional data was periodically loaded into a separate analytics database where often long-running OLAP queries could be run without slowing the transactional database.
Hybrid Transactional/Analytical Processing (HTAP) with In-Memory Computing
The GridGain in-memory computing platform provides speed and scale to systems which integrate it into their architecture. By moving to an in-memory platform which processes queries up to 1,000,000x faster than disk-based systems, concerns about analytic processing slowing the transactional database are no longer an issue. Both transactions and analytics can be run on the same GridGain in-memory dataset with no negative impact on performance. In addition, the amount of data that can be held in-memory can be scaled out simply by adding commodity nodes to the GridGain cluster, allowing organizations to keep an arbitrarily large set of data in-memory to support analytical processing requirements.
The GridGain in-memory computing platform is also ANSI SQL-99 compliant, allowing users to easily run SQL analytical queries against the in-memory transactional dataset. Full SQL indexing and querying at in-memory speeds provides a powerful and fast solution for ad hoc or scheduled HTAP analytical jobs.