ECE 697: Real-Time Systems

Instructor: C. M. Krishna» krishna@ecs.umass.edu; (413) 545-0766» Office Hours (on-campus): Tues: 4:00--4:45 PM

Off-campus Contact: By e-mail and telephone Course Grading:

» Three in-term tests: 18% each» Final exam (cumulative): 26%» Homework/simulation exercises: 20%

Homework is to be done individually

Coverage

Introduction to real-time systems Performance measures Task allocation and scheduling techniques Power and energy issues Communication algorithms Fault tolerance and reliability evaluation Clock synchronization

Textbook

» C. M. Krishna and K. G. Shin, Real-Time Systems, McGraw-Hill, 1997.

» On-campus students: I’ve put a copy on reserve in the Physical Sciences Library; I’ll make another copy available in the Architecture & Real-Time Lab

» Off-campus students: Check your nearest technical library

» Also available from bookstores (online and traditional)» See the course web page for a pointer to the typo list:

http://www-unix.ecs.umass.edu/~krishna/rtcourse.html

Course Notes

A few PowerPoint slides: Will be posted on the course website

Mostly handwritten in class: The more important bits will be scanned and available through the course website after the lecture

Readings beyond the text will be used for » Real-time communication protocols» Energy- and power-aware computing

Today’s Topics

What is a real-time system?» General characteristics» Hard and soft real-time systems

Performance Measures» Why are they important?» For general-purpose systems» For real-time systems

Uniprocessor task scheduling

What is a Real-Time System?

Any system in which a deadline plays a central role in its perceived performance» But timely response is important for general-purpose

systems, too!» There is no hard-and-fast demarcation between a real-

time system and a general-purpose system» Systems in the control loop are always real-time

Types of RTS

Hard Real-Time Systems» Missing a deadline (or series of deadlines) can cause a

significant loss to the application.» Examples: Fly-by-wire, power-plant, and grid control

Soft Real-Time Systems» Missing a deadline causes the quality of service to

degrade, but nothing terrible happens» Examples: Video-on-demand, teleconferencing

Example: Fly-by-wire

Used initially in military aircraft» Dynamics time-constants are too small for humans to

be effective controllers» Philosophy:

Pilot sets policy Computer carries out low-level actions to implement that

policy» If too many deadlines are missed in a row, the aircraft

can crash

Feedback Loop

(From C. M. Krishna & K. G. Shin: NASA Con. Report 3807, 1984)

Impact of Feedback Delay (Simulation Example)

Elevator Deflections During Landing

Performance Measures

Traditional Measures» Throughput: Average number of instructions processed

per second» Availability: Fraction of time for which the system is

up» Reliability: Probability that the system will remain up

throughout a designated interval

Special-Purpose Measure

Performability» Published by John Meyer in 1980» Identify accomplishment levels, {A0, A1, A2, …, An},

for the application» Determine the probability, P(Ai), that the real-time

system will be able to perform in such a way that Ai will be accomplished

» Performability is the vector (P(A0), P(A1), …, P(An))» Application-focused measure

Task Allocation and Scheduling

How to assign tasks to processors and to schedule them in such a way that deadlines are met

Our initial focus: uniprocessor task scheduling

Uniprocessor Task Scheduling

Initial Assumptions:» Each task is periodic» Periods of different tasks may be different» Worst-case task execution times are known» Relative deadline of a task is equal to its period» No dependencies between tasks: they are independent» Only resource constraint considered is execution time» No critical sections» Preemption costs are negligible» Tasks must be completed for output to have any value

Standard Scheduling Algorithms

Rate-Monotonic (RM) Algorithm:» Static priority» Higher-frequency tasks have higher priority

Earliest-Deadline First (EDF) Algorithm:» Dynamic priority» Task with the earliest absolute deadline has highest

priority

RMA

Task priority is inversely proportional to the task period (directly proportional to task frequency)

At any moment, the processor is either» idle if there are no tasks to run, or» running the highest-priority task available

A lower-priority task can suffer many preemptions To a task, lower-priority tasks are effectively

invisible

RMA

Example Schedulability criteria:

» Sufficiency condition (Liu & Layland, 1973)» Necessary & sufficient conditions (Joseph & Pandya,

1986; Lehoczky, Sha, Ding 1989)

RMA

Critical Instant of a Task: An instant at which a request for that task will have the largest response time

Critical Time-zone of a Task: Interval between a critical instant of that task and the completion time of that task

Critical Instant Theorem: Critical instant of a task T_i occurs whenever T_i arrives simultaneously with all higher-priority tasks

RMA: Scheulability Check

The Critical Instant Theorem leads to a schedulability check:» If a task is released at the same time as all of the tasks

of higher priority and it meets its deadline, then it will meet its deadline under all circumstances

RMA: Schedulability Test

If a task is released simultaneously with all higher-priority tasks, determine when it will be done

If this completion time is no later than this task’s deadline, we have succeeded with this task

Find a systematic procedure to turn this process into a necessary-and-sufficient schedulability check

RMA: Schedulability

Start with a single-task set and obtain its schedulability conditions

Extend this to a two-task set Exploit any intuition gained to generalize this