Lecture 7 Process Management and Single Processor Scheduling

Outline

1. Real-time OS

• Definition of real-time OS
• Real-time tasks

2. Real-time Scheduling
3. Priority Inversion

Definition of real-time OS

• Definition of real-time OS
  • An operating system that depends on both time and functionalities for correctness.
• Performance metrics for real-time OS
  • Timeliness of time-constrained tasks (deadlines)
  • Speed and average performance are relatively unimportant
• Characteristics of real-time OS
  • Predictability of time-constrained tasks

Classification of Real-Time OS

• Strong/hard real-time OS
  • Requires critical tasks to be completed within a specified time.
• Weak/soft real-time OS
  • Processes with high priority should be completed as soon as possible, but not necessarily required.

Real-time tasks

• Task (Work Unit)
  • A single computation, file read, message transfer, etc.
• Task properties
  • Resources required to complete the task.
  • Deadline parameter

Periodic real-time tasks

• Periodic real-time tasks: a series of similar tasks
  • Tasks are repeated regularly.
  • Period p = task request interval (0 <p)
  • Execution time e = maximum execution time (0 < e < p)
  • Utilization U = e/p
    
• Schedulable: if $\sum_{p_i} \frac{e_i}{p_i}\leq 1$; otherwise non-schedulable

Soft Deadline and Hard Deadline

• Hard deadline
  • Missing the task deadline will result in catastrophic and very serious consequences.
  • It must be verified that the deadline can be met in the worst-case scenario.
• Soft deadline
  • The task deadline is usually met.
  • If it cannot be met, lower the requirement.
  • Do your best to meet the deadline.

Schedulability

• Schedulability refers to the ability of a real-time OS to meet task deadlines.
  • The execution order of real-time tasks needs to be determined.
  • Static priority scheduling: Task priority does not change during execution.
  • Dynamic priority scheduling: Task priority can change during execution.

Outline

1. Real-time OS

2. Real-time scheduling

• Rate Monotonic(RM) Algorithm
• Earliest Deadline First(EDF) Algorithm
• Least Laxity First(LLF) Algorithm

3. Priority Inversion

Real-time scheduling

• Static Priority Scheduling: Rate Monotonic(RM) Algorithm

  • Priorities are assigned based on the task periods
  • The shorter the period, the higher the priority
  • Execute the task with the shortest period

• Dynamic Priority Scheduling: Earliest Deadline First(EDF) Algorithm

  • The task with the earliest deadline has the highest priority
  • The task closest to the deadline is executed first

If there is shared resource occupation between tasks, tasks with high priority may be delayed!

Rate Monotonic(RM) Algorithm

• Determine the task priority based on task periods (the shorter the period, the higher the priority, preemptive)

• Process P1: e=20 p=50

• Process P2: e=35 p=100

Rate Monotonic(RM) Algorithm

• Determine the task priority based on task periods (the shorter the period, the higher the priority, preemptive)

• Process P1: e=25 p=50

• Process P2: e=35 p=80

Earliest Deadline First(EDF) Algorithm

• Issues of Fixed priority: some tasks may miss the deadline

• Process P1: e=10 p=20

• Process P2: e=25 p=50

Earliest Deadline First(EDF) Algorithm

• Issues of Fixed priority: some tasks may miss the deadline

• Process P1: e=10 p=20

• Process P2: e=25 p=50

Earliest Deadline First(EDF) Algorithm

• Issues of Fixed priority: some tasks may miss the deadline

• Process P1: e=10 p=20

• Process P2: e=25 p=50

Earliest Deadline First Algorithm (EDF, Earliest Deadline First)

• Priorities of tasks are dynamically assigned based on their deadlines. The closer the deadline, the higher the priority.
• Process P1: e=10 p=20
• Process P2: e=25 p=50

Lowest Laxity First Algorithm (LLF)

• Determine task priority according to task urgency or laxity level

• The higher the urgency, the higher the priority.

• Laxity = Deadline - Execution Time - Current Time

• Process P1: e=10 p=20; Process P2: e=25 p=50

Outline

1. Real-time OS
2. Real-time Scheduling

3. Priority Inversion

• priority inheritance
• Priority Ceiling Protocol

Priority Inversion

The phenomenon of a high-priority process waiting for a resource occupied by a low-priority process for a long time.

• Priority inversion problem exists in priority-based preemptive scheduling algorithms. Priority: T1>T2>T3

Priority Inversion

The phenomenon of a high-priority process waiting for a resource occupied by a low-priority process for a long time.

• Priority inversion problem exists in priority-based preemptive scheduling algorithms.

Priority: T1>T2>T3

Priority Inheritance

• A low-priority process that is holding a resource inherits the priority of a high-priority process that requests the same resource.
• The priority of the resource-holding process is raised only when the low-priority process is blocked.
  

Priority Inheritance

• A low-priority process that is holding a resource inherits the priority of a high-priority process that requests the same resource.
• The priority of the resource-holding process is raised only when the low-priority process is blocked.
  
  Note: Critical section refers to a code segment that accesses shared resources in a mutually exclusive manner.

Priority Inheritance

• A low-priority process that is holding a resource inherits the priority of a high-priority process that requests the same resource.
• The priority of the resource-holding process is raised only when the low-priority process is blocked.
  
  Note: Critical section refers to a code segment that accesses shared resources in a mutually exclusive manner.

Priority Inheritance

• A low-priority process that is holding a resource inherits the priority of a high-priority process that requests the same resource.
• The priority of the resource-holding process is raised only when the low-priority process is blocked.
  
  Note: Critical section refers to a code segment that accesses shared resources in a mutually exclusive manner.

Priority Inheritance

• A low-priority process that is holding a resource inherits the priority of a high-priority process that requests the same resource.
• The priority of the resource-holding process is raised only when the low-priority process is blocked.

Note: Critical section refers to a code segment that accesses shared resources in a mutually exclusive manner.

Priority Inheritance

• A low-priority process that is holding a resource inherits the priority of a high-priority process that requests the same resource.
• The priority of the resource-holding process is raised only when the low-priority process is blocked.

Note: Critical section refers to a code segment that accesses shared resources in a mutually exclusive manner.

Priority Inheritance

• A low-priority process that is holding a resource inherits the priority of a high-priority process that requests the same resource.
• The priority of the resource-holding process is raised only when the low-priority process is blocked.

Note: Critical section refers to a code segment that accesses shared resources in a mutually exclusive manner.

Priority ceiling protocol

• The priority of a process holding a resource is set equal to the highest priority of any process that might potentially request that resource.
  • The priority of the resource-holding process is elevated regardless of whether a wait occurs or not.
  • The priority of the resource-holding process is higher than the upper limit of the priority of all locked resources in the system, thereby preventing blocking of task execution within a critical section.

Summary

1. Real-time OS

• Definition of real-time OS, real-time tasks

2. Real-time scheduling

• Rate Monotonic(RM) Algorithm, Earliest Deadline First(EDF) Algorithm, Least Laxity First(LLF) Algorithm

3. Priority Inversion

• Priority Inheritance, Priority Ceiling Protocol