The main information systems used for operational support in a business are transaction processing systems. This type of system processes data about transactions, which are events that have occurred that affect the business, such as the sale or purchase of goods.
A transaction processing system has 3 main purposes: keep records about the state of an organization, process transactions that affect these records, and produce outputs that report on transactions that have occurred. For example, an inventory control system tracks records about inventory, processes sales and purchases of inventory, and produces reports about the amount and value of items on hand, on order, etc.
Transaction processing systems exist in all areas of an organization, and in all types of organizations. TPSs can be used by employees (order entry) or customers (bank ATMs). They can use various types of hardware, software, and networks. TPSs use stored data in both files and databases, and many types of programs.
TPS Structure and Functions
The structure of a TPS generally consists of a user or other TPS that interacts with TPS software, and stored data used by the TPS. Users of the system are typically personnel who work with business transactions, such as salespeople. Input data comes from users and other TPS. Output includes data back to other TPSs, screens, and reports.
Like other information systems, TPSs perform four main functions: input, processing, output, and data storage. The input function accepts data for processing from outside the system. The processing output transforms the data in some way. The output function makes the processing results available outside the system. The storage function stores the data for use.
Before transaction data can be brought into a TPS, it must be acquired from its source. This step is called data capture. The receiving report is an example of a source document that is used to perform data entry. Data entry means using a screen similar to the one in this figure to enter data. Once the data is entered, a program checks it for errors, a process called data validation. For example, inventory receiving data might be checked to ensure that all number entries are numeric that the item is known to the system, and that quantities are in acceptable ranges.
Much output from a TPS is in the form of reports. Several types of reports are commonly produced by TPSs. A detailed report lists details about transactions, such as inventory movement. A summary report summarizes data at various levels. Exception reports indicate data that are exceptions to some condition or standard.
Data in a TPS is stored to data files and databases. Two types of stored data are commonly found in TPSs: transaction and master data. Master data is the main data used by the TPS. Master data is usually permanent data that remains in the system as long as the system is in use. For example, an inventory system would have an inventory master file with one record for each item in inventory. Transaction data is data about transactions that have occurred. Transaction data usually remains with the system only until the transactions are processed. The transaction data is then replaced with other transaction data for new transactions. For example, in the inventory control system, a transaction file would contain data about inventory transactions. That data would be used to update the inventory master file. Once the update was complete, the transaction file would be cleared and a new list of transactions would be started.
After the data is created, the TPS can retrieve data from the file or database, a process called accessing the data. The data is accessed to produce reports, update master files, and other purposes. Sometimes, before a TPS can access data, it must be arranged in an order that is useful to the system, a process called sorting. For example, the inventory system may produce a report of the 5 most expensive items. The item list should then be sorted in descending order of cost.
The data put into a file or database when it is created will become obsolete over time. The TPS must update the data periodically to keep it current. Updating can involve adding new data, changing current data, and deleting old data. Transaction data is used to update master data.
Processing involves manipulating data within the system. One function that just involves data processing and not any of the other functions is computation, such as calculating details and totals. Another processing function is decision making, or checking for conditions and acting on them.
Data in a TPS can be processed using two basic approaches: batch and online processing. A TPS may use both online and batch processing. In batch processing, data for all transactions to be processed are prepared in form understandable to the computer before actual processing begins. Then the batch of data is processed by the computer, and the resulting output is received in a batch. An example of batch processing is overnight payroll check processing and printing.
With online processing, or OLTP, a person uses a screen and keyboard or I/O device connected to the computer at the time the processing is done. Each set of data for a transaction is entered directly into the computer. The data is processed, and the output is received before the next input data is entered. Online processing may also be called interactive processing. An example of online processing is an airline reservation system.
The term real-time may also be used to indicate online processing. Real-time means the processing is performed immediately after the data is entered. This description is not quite accurate for online processing. If there are many users of an online system, processing may not begin for some time after the input is received. The amount of time depends on the hardware, software, and number of users. One example of a real-time system is a manufacturing monitoring system. If cars are being manufactured and theres a problem with the computerized assembly line, the system alerts operators immediately.
TPSs must have procedures to ensure the completeness of the data processing, and to minimize the chance of errors. In general, these procedures are called controls. Many types of controls are used, including control totals, audit trails, and backup and recovery procedures.
All data in a TPS may not be processed for various reasons, including hardware and software failures, and human error. One way that a TPS checks that all data is processed is through control totals. A control total is a number that is computed when data enters a system, and again after the system has processed the data. For example, in the inventory control system, a number reflecting an initial count of receiving reports to be entered would be entered into the system. After the receiving reports are entered, a count of the documents entered would be produced. There are also other types of control totals. Document counts are only one type.
An audit trail is a way of tracing the effect of data through a system. A good audit trail is one in which someone can start with the output and go back through the system to the source document and vice versa. For example, the inventory receiving system could produce a report of all item counts. This could be used to trace each item received back to its receiving report source document.
Computer systems sometimes fail, causing a loss of corruption of data. The failure may be caused by a malfunction in the hardware or software, or because of something outside the system, such as a lightning strike. The main way of ensuring against loss of data is to use a backup procedure. This means copying important data and programs to a portable media and moving it somewhere away from the main computing site. Important files and databases are usually backed up daily or weekly.
The backup of the stored data can be copied back to the system, a process known are restoring the data. If a system fails, recovery procedures tell the computer professionals how to use the backups to recover from a crash.