Course contentsShow
← Computer Science
Lesson 1258 of 2,87229. CPU SchedulingFree lesson

Priority Inversion and Aging

Understand the starvation problem in priority scheduling and how aging gradually increases waiting process priorities.

Priority Inversion and Aging

You'll learn how priority scheduling can accidentally starve low-priority processes, and how aging fixes this by gradually boosting their priority over time.


The Starvation Problem

In priority scheduling, the CPU always runs the highest-priority process first. Sounds fair, right? But here's the catch: if high-priority processes keep arriving, low-priority processes may never get CPU time. They sit in the ready queue indefinitely, waiting forever—a problem called starvation.

Analogy: Imagine an emergency room where critical patients (high priority) are always treated first. If emergencies keep arriving non-stop, someone with a minor injury might wait for days, even though their issue deserves eventual care.

What is Aging?

Aging is a simple, elegant solution: the longer a process waits in the ready queue, the more we gradually increase its priority. Eventually, even the lowest-priority process becomes high enough to get scheduled.

For example:

  • A process starts at priority 5 (low)
  • Every 10 minutes it waits, we boost its priority by 1
  • After 50 minutes: priority becomes 10 (high enough to run)

This ensures fairness—every process eventually gets CPU time, preventing indefinite starvation.

Priority Inversion (Brief Context)

Priority inversion is a related but different problem: a high-priority process waits for a resource held by a low-priority process, while a medium-priority process runs, delaying everyone. Aging doesn't solve this directly, but understanding both helps you see the complexities of priority-based scheduling.


Key Takeaway: Aging prevents starvation in priority scheduling by gradually increasing the priority of waiting processes, ensuring all processes eventually receive CPU time regardless of their initial priority.