Online transaction processing (OLTP), also known as real-time transaction processing, refers to a usage paradigm of database systems and business applications in which the processing of transactions takes place directly and promptly, i.e. without significant time delay.
The counterpart is batch processing, in which all business transactions are collected and processed in batch runs – often at night, free of online operation. The naming dates back to the 1960s/1970s because the data (and often also programs) were available as punch cards and were read in and processed as stacks. With this designation, the aspect also plays a role that the tasks (as with the punch card stack) can only be carried out one after the other, while with OLTP many different business operations take place simultaneously.
The main technical focus of OLTP is on transaction security for parallel requests and changes, on minimizing the response time of requests and on the highest possible throughput (number of transactions per unit of time). The efficiency of OLTP systems depends on the selection of suitable hardware (database server, network components such as LAN and WAN) and software (database management system). By observing transaction criteria (see ACID), it is ensured that the consistency of the database (there can be several distributed ones nowadays) is maintained and that data is never stored incompletely or inconsistently.
---
OLTP database systems typically store the transactions of a business transaction in the highest level of detail required for the company’s operations. About the term OLTP, one also encounters the term Online Analytical Processing (OLAP), which deals with the aggregation, preparation and evaluation of the operational data for a longer period.
The classic use case of OLTP is the computer-aided execution of the business processes of companies that make up the day-to-day operational business. This task is performed, among other things, by so-called ERP systems. This is where the master data such as personnel or inventory and business transactions such as purchasing, sales, space postings, account movements, etc. are processed. Modern companies are largely dependent on the existence and correctness of this data, as it reflects all resources such as personnel, finances, goods and/or services together with the current changes caused by business activities. Therefore, such applications are classified as mission-critical.
Database systems have the task of backing up all data in such a way that no data loss can occur and the data itself remains consistent at all times – even after a partial loss. This requirement is more stringent for application systems, which since the 1990s can also be implemented on distributed databases that run on different servers and/or in different locations – also connected worldwide via WAN. Other applications include content management, knowledge bases, webshops, directory services, etc.
