Real-time is a class of data processing that responds to events and demands as they occur. The speed of real-time processing depends on the problem space. It is common for seconds or minutes to be considered real-time. Some problems demand that processing complete in milliseconds or microseconds to be considered real-time. The following are common levels of real-time processing.
Hard
Missing a deadline is a total failure. For example, an self-driving car that doesn't apply brakes in time to avoid hitting a pedestrian.
Firm
Missing a deadline means that a result is worthless but the system as a whole is able to tolerate a degree of failure. For example, sensor processing in a self-driving car may fail to meet a deadline but the car may continue to operate normally. The car may be designed with more sensors than it needs such that a failure of a single sensor reading doesn't cause the car to make poor decisions. In theory, a high percentage of sensor readings could fail and the car could continue to be safe due to reliability engineering techniques such as sensor competition.
Soft
A result is still useful when a deadline is missed. For example, an online credit card application promises customers an instant answer. If the response isn't completed in one minute many customers will close the page. The result is still useful as it can be emailed to the customer.
Cookies help us deliver our services. You have choices regarding these cookies. Please visit our privacy policy, cookie policy and consent tool to learn more.
Copyright 2002-2026 Simplicable. All rights reserved. This material may not be published, broadcast, rewritten, redistributed or translated. Report violations here.