Online transaction processing
Online transaction processing ( OLTP ), online transaction processing , also real-time transaction processing , describes a usage paradigm of database systems and business applications in which the processing of transactions takes place directly and promptly, i.e. without any significant time delay.
The counterpart is batch processing , in which all business transactions are collected and processed in batch runs that are often nightly and free from online operations. The naming ( English batch 'stack') comes from the 1960s / 1970s, because the data (and often also programs) were available as punch cards and were read and processed as a stack. With this name, the aspect also plays a role that the tasks (as with the pile of punched cards) can only be carried out one after the other, while with the OLTP many different business operations take place at the same time.
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 time unit). 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 (nowadays there can also be several distributed ones) is maintained and that data is never stored incompletely or inconsistently.
OLTP database systems typically store the transactions of a business process with the highest level of detail that is required for the company's operational activities. Based on the term OLTP, one also encounters the term Online Analytical Processing (OLAP) , which deals with the aggregation, processing and evaluation of operational data over a longer period of time.
application
The classic application of OLTP is the EDP-supported implementation of the business processes of companies that make up the daily 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 processes such as purchasing, sales, space bookings, account movements, etc. are processed. Modern companies are largely dependent on the existence and correctness of this data, since all resources such as personnel, finances, goods and / or services are shown in it, together with the current changes caused by the business activity. Therefore, such applications are classified as business critical.
The task of database systems here is to back 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 in the case of application systems which, since the 1990s, have also been able to be implemented on distributed databases that run on different servers and / or at different locations - including those connected worldwide via WAN.
Other applications are content management , knowledge databases , web shops , directory services, etc.
