Data processing:Processing methods

Processing methods

In DP systems there are two processing methods: batch processing and real-time processing.

In batch processing, the transaction data is accumulated over the course of several hours onto transaction files and then processed as a batch. This type of processing has the advantage that batch 'runs' can be scheduled for times in the day when the computer is not heavily used, leaving peak periods clear for work that cannot be handled in this way. However, because the transaction data is not processed immediately, this method suffers from the disadvantage that the master data, such as stock balances, are normally slightly out of date.

In real-time processing the computer deals with the trans­action data immediately it arises; the data is not first recorded on a transaction file for later processing. Instead, the master file is updated immediately and any transaction documents are produced there and then. However, the data is recorded on a transaction file after the processing has taken place, the sole purpose of this file being to print out an audit trail at the end of the day (see page 217).

Real-time processing has the advantage that the master files are always up to date. The disadvantage is that it makes heavy use of the computer during busy periods. The steadily increasing power of computers is lessening the importance of this disadvantage, with the result that there has been a shift in recent years away from batch processing towards real-time processing.

Batch processing is the traditional way of processing data by computer, involving a large data processing department. Data is sent by the sales, purchasing, and other user departments in the DP department on source documents of various sorts. Working from these, the data entry clerks in the DP department key the data onto transaction files for later processing (see page 209). The output, when it is produced, is dealt with by data control clerks and distributed to the user departments.

Real-time processing, in contrast, passes these tasks over to the user departments. VDUs are installed in the depart­ments, and staff key in the data as it arises and receive an immediate response on their screens. Transaction documents may be printed out immediately on printers in the user departments.

Accessing methods

The process of storing and retrieving data is called data access, and a number of data access methods are used in computer systems. Data on transaction files is always accessed in the order in which it was entered into the file, i.e. chronologically. This is called serial access. Serial access is similar to sequential access, referred to on page 162: the computer begins at the beginning and reads through and processes each item of data until it reaches the end of the file.

(Note that the term sequential access is applied when records are written to and read from the file in key sequence, i.e. code number sequence if the code number is the key. Records on master files are often accessed sequentially. See page 130 for information on key fields.)

Serial and sequential access methods are used in DP applications where all the records in a file have to be accessed. An example in the case of master files is the production of monthly customer statements. Every custom­er's account details held on the customer file must be looked up in order to produce the statements. If a file is accessed only serially or sequentially, then it can be stored on magnetic tape; this has to be read in sequence from begin­ning to end, but it is a cheaper form of storage than disk. However, files are often accessed in a random way (see below), and then disk storage is essential.

In many applications, records on master files have to be accessed in a random manner. The access device in this case goes straight to the spot on the disk storage device (normally magnetic disk) where the record is held. Invoicing is one example of this: to produce an invoice a particular customer has to be looked up in the customer file, and particular stock lines have to be looked up in the stock file.

A common random access method is indexed sequential access. In this, the records are stored in key sequence in the file, with an index which catalogues each record's location and thus allows the computer to find it immediately. In this type of storage method the records can also be read sequen­tially, if this is necessary.

Comments

Popular posts from this blog

WORKED EXAMPLES ON PROCESS SPECIFICATION.

Why do we need information systems, management structure, requirements of information at different levels of management.

The User Interface:Establishing User Interfaces