A defining requirement of an online real-time computing system is the ability to process inputs and produce outputs without unacceptable delay. Which option captures this requirement best?

Introduction:
Real-time systems are designed to react to events under strict timing constraints. “Online, real-time” emphasizes that data arrives continuously and must be acted on promptly, as in control loops, avionics, medical devices, or trading systems.

Given Data / Assumptions:

• “Real-time” implies meeting deadlines, not merely high speed.
• “Online” implies immediate availability of inputs and outputs during operation.
• We seek the essential requirement, not an implementation detail.

Concept / Approach:
The core requirement is deterministic, bounded latency from input to response. Whether a system uses one CPU or many is an implementation choice, not a definition. Offline batch processing contradicts “online real-time.”

Step-by-Step Solution:

Verification / Alternative check:
In schedulability theory (e.g., rate-monotonic, earliest-deadline-first), the focus is on meeting timing constraints, not CPU count.

Why Other Options Are Wrong:

• Multiple CPUs mandatory: real-time systems can be single-CPU; determinism matters.
• Offline batch processing: contradicts online, real-time behavior.
• All of the above: not possible because one choice contradicts the definition.
• None of the above: one choice is correct.

Common Pitfalls:
Equating high throughput with real-time; assuming parallelism guarantees deadlines; ignoring jitter and worst-case latency.

Final Answer:
Deterministic response with no unacceptable processing delay.