What Is a Real-Time System?

A real-time system is any information processing system that performs application functions within predictable and specific time constraints, usually by an order of microseconds. Real-time systems are designed to support mixed-criticality workloads, where basic functions that can tolerate delays, such as displaying data on a monitor, execute concurrently with critical functions, such as controlling a robotic arm on an assembly line.

To qualify as a real-time system, a computer or device must satisfy two requirements:

• Timeliness: The computer must be able to produce an expected result by a specific deadline.
• Time synchronization: The computer must be able to coordinate its internal clock with other computers so they can operate in unison.

How Real-Time Systems Work

Real-time systems use various techniques to prioritize critical workloads over noncritical workloads. These techniques can include reserving access to key system resources, such as processor cache or memory, for important or time-sensitive tasks. The goal of real-time systems is to achieve predictability, also known as deterministic performance. Latency, jitter, and precision are also important factors for understanding predictability:

• Latency: Latency is the measurement of time between two events.
• Jitter: Jitter refers to the difference in latency between iterations of the same event.
• Precision: Precision is a measure of variation in task completion time across multiple components.

Time-Sensitive Networking (TSN)

Multiple real-time systems working together require standardized connectivity protocols to ensure smooth operation (consider the need to coordinate several conveyor belts and robotic arms on a single production line). For devices connected over Ethernet, this set of standards is known as Time-Sensitive Networking (TSN).

Why Real-Time Systems Are Important

Real-time processing is important anytime the failure to deliver predictability could result in negative outcomes. For example, in media and entertainment, broadcasters may have service-level agreements (SLAs) to deliver live content within a specific time frame, and failure can incur financial penalties or loss of credibility with audiences. By contrast, in automotive use cases, a delayed signal between a vehicle’s controls and its engine components could lead to a car accident and subsequent injuries.

Benefits of Real-Time Systems

In addition to being a requirement for specific technology applications, real-time systems offer numerous benefits:

• AI adoption: Real-time systems help support artificial intelligence (AI) and automation across all industries by supplying AI with the most up-to-date information—which can include data about the status of real-time market positions in financial trading or visual inputs for an AI defect detection system on a factory production line.
• Smooth coordination: Real-time systems are designed to perform tasks that must be executed within precise cycle deadlines, down to the microsecond, and often in coordination with other devices with synchronized internal clocks.
• Predictability and reliability: Because real-time systems process data in predictable time frames, execution of tasks or workloads is practically guaranteed within a measurable threshold. This provides more visibility into system functionality and more consistent results.
• Prioritization and isolation: Generally, real-time systems provide solution designers with more tools and techniques to control resource allocation at both the hardware and software levels. With these tools, designers can more easily prioritize or even isolate key processes from others in mixed-criticality workloads.

Challenges of Deploying Real-Time Systems

Specialized hardware and software are required to enable real-time systems. This includes processors, firmware, operating systems, and hypervisors that support task prioritization, TSN, and other features to help manage resource contention. Businesses also need the expertise to configure software and BIOS settings for real-time operation.

Depending on the scale of a deployment, real-time systems can also be more difficult to change or update. For example, in a factory environment with numerous conveyor belts and robots all working in tandem, changes to the device topology can incur extreme costs and an extensive calibration effort across the entire plant. For this reason, manufacturers will typically avoid upgrading their real-time infrastructure until there’s a high return-on-investment (ROI) opportunity.

Soft Real-Time Systems vs. Hard Real-Time Systems

Real-time systems vary by degrees of delay tolerance, where a missed deadline results in consequences for the systems or the people using them.

In a soft real-time system, computers or equipment will continue to function after a missed deadline but may produce a lower-quality output. For example, latency in online video games can impact player interactions, but otherwise present no serious consequences.

Firm real-time systems have less delay tolerance but may still function, although the output may be unusable. Examples include global positioning systems (GPS) or stock trading platforms, where outdated information is simply not helpful for the end user.

Hard real-time systems have zero delay tolerance, and delayed signals can result in total failure or present immediate danger to users. Flight control systems and pacemakers are both examples where timeliness is not only essential but the lack of it can result in a life-or-death situation.

Examples of Real-Time Systems

Real-time requirements are present to varying degrees in countless applications, resulting in many different technology form factors:

• Wearable technology: Wearables and patient monitoring devices rely on real-time processing to enable telehealth and remote care.
• Embedded computers: Real-time systems help support AI in vehicles and responsive flight controls in aircraft.
• Edge devices and servers: Many industrial, logistics, or transportation use cases leverage edge computing for real-time processing, often enhanced with AI to support automation.
• Data center servers: Real-time-capable server racks and clusters can process, analyze, or transmit data at scale, typically in regard to real-world events as they happen, such as live sports broadcasting or financial markets trading.
• Public cloud resources: Some cloud service providers (CSPs) offer instance types specifically for time-sensitive workloads, such as data streaming for content delivery.

Real-Time System Use Cases

The need for real-time systems is especially crucial for businesses that depend on robotics, avionics, and medical devices. Businesses engaged in high-precision activities, such as the energy sector, also rely on real-time processing for safety and productivity.

Process Control Systems

Process control systems enable workers to monitor and control heavy machinery, often in environments where production is continuous and downtime is costly. Real-time systems help keep automated equipment running smoothly and provide timely status updates to equipment operators.

Machine Vision

AI machine vision enables computers to make decisions based on visual inputs. Real-time systems help ensure machine vision–enabled computers can process and act on what they see when they see it.

Robotics

Robotics technologies enable machines to interface with real-world objects, people, or environments, which requires precise timing to help ensure safety.

Manufacturing

Manufacturers that embrace Industry 4.0 transformation use advanced technology, including real-time systems, machine vision, and robotics, to further their business and sustainability goals.

Healthcare and Patient Monitoring

In healthcare settings, real-time systems in devices such as heart rate monitors are critical to keeping clinicians informed about patient health from moment to moment while allowing them to respond to emergencies quickly.

The Future of Real-Time Systems

Many real-time-enabled technologies that were considered futuristic in the past—such as robotics, remote patient care, live media streaming, and self-driving vehicles—are available now. Some of these innovations are already widespread or on their way to mainstream adoption. Real-time systems will continue to become more sophisticated in function but simple to deploy, strengthening the foundation for all manner of embedded, edge, and wearable technology.