Data transactions (OLTP) and analytics (OLAP) have long been treated as very different concerns. Analyzing high volume transactional data traditionally required complex and hard to maintain ELT / ETL integration batches that ran overnight, causing any insights to be based on data that is already outdated.
What if we could transact data very fast, on an open-source horizontally high scalable NoSQL system, and that data be automatically and constantly written to a massive parallel analytical database - allowing near real-time transactions and analytics?
What if we could cache back on the transactional system any analytical data insights or machine learning algorithm results, pushing those analytical findings back to the applications, allowing real closed-loop analytics?
This talk introduces an open-source solution that integrates the fastest scalable, highly available and fully consistent In-Memory Data Grid (Apache Geode / GemFire) to the first open-source massive parallel data warehouse (Greenplum Database) in a hybrid transactional and analytical architecture that is extremely fast, horizontally scalable, highly resilient and open source.
The session also features a live demo, showing a real case of real-time closed-loop analytics and machine learning using the featured solution.