The transactions are entered in batches by keyboard and stored in transaction files. These batches consist of thirty or so records which are given a batch control ID. The batches are then run through a validation process and to make sure the batches balance a computed total is compared with a manually produced total. This helps to ensure that all data is entered without error or omission. The actual updating of master files only takes place after verification and validation are complete. This means batch processing is often run overnight, unattended. A new master file is produced as a result of a batch processing run. The original master file is kept along with a previous version.
After processing the output is produced, and is usually printed media such as payslips or invoices, although this is changing with the advent of the web.
There is really no such thing as real-time. The best you will get is a few milliseconds from input to response. However such fast systems are used in critical systems that control aircraft or the manufacture of sensitive or dangerous compounds.
Online processing means users directly enter information online (usually, online, in this case, means online to a central processor, rather than its modern connotation of the internet, but it could mean both!), it is validated and updated directly onto the master file. No new file is created in this case. Therefore, there is near immediate input process, and output. Imagine a cash dispenser transaction or booking a holiday at a travel agents or over the Internet. Compared with batch processing the number of transactions will be few.
Online comes in many different flavours such as centralised, distributed, time-share etc and the choice of architecture will depend on cost, speed needed, number of users, number of transactions and time needed for a response.
Chat with our AI personalities