The information technology field is full of obfuscation. There are many three letter acronyms, including the meta three letter acronym TLA. Now in the era of Big Data and Cloud, we have four letter acronyms, what I call FLAS, two of which are HTAP and ACID. People have asked me what these are, so here is an explanation.
ACID has a new meaning. No longer is it just Atomicity, Consistency, Isolation, and Durability. It can also mean Algorithms, Cloud, IoT, and Data. HTAP is Hybrid Transactional Analytic Processing. Transactional processing systems are by default not optimized for analytical processing. Transactional processing works well when data is stored in indexed rows. That type of storage does not work well for analytic processing, which involves the aggregation of columns of data. By arranging and storing data in columns, aggregations can be done much faster. A tried and true technique for providing fast analytical processing and fast transactional processing is to have ETL or ELT processes to periodically build columnar OLAP storage from row oriented OLTP storage. The disadvantage is the latency introduced.
HTAP solutions provide both OLAP and OLTP storage allowing analytical processing on current transactional data without slowing down the transactional processing.
Of course I’m writing about FLAS because my job title as of this writing is DPSA, Data Platform Solution Architect.