CHAPTER 3 A Reference Model of Real-Time Systems

Typical Real-Time Applications controller reference A/D control-law computation input D/A A/D Sensor Plant Actuator Fig. 1-1 A digital controller © Cheol-Hoon Lee SSLab, CNU, 2015 Ch.3 A Reference Model of Real-Time Systems

Typical Real-Time Applications - continued Do the following in each 1/180-second cycle: Validate sensor data and select data source; in the presence of failure, reconfigure the system. Do the following 30-Hz avionics tasks, each once every six cycles: - Keyboard input and mode selection - data normalization and coordinate transformation - tracking reference update Do the following 30-Hz computations, each once every six cycles: - control laws of the outer pitch-control loop - control laws of the outer roll-control loop - control laws of the outer yaw- and collective control loop Do each of the following 90-Hz computations once every two cycles, using outputs produced by 30-Hz computations and avionics tasks as input: - control laws of the inner pitch-control loop - control laws of the inner roll- and collective-control loop Compute the control laws of the inner yaw-control loop, using outputs produced by 90-Hz control-law computations as input Output commands. Carry out built-in-test. Wait until the beginning of the next cycle. Fig. 1-3 An example: Software control structure of a flight controller © Cheol-Hoon Lee SSLab, CNU, 2015 Ch.3 A Reference Model of Real-Time Systems

Ch.3 A Reference Model of Real-Time Systems In This Chapter A reference model of real-time systems a workload model that describes the applications a resource model that describes the system resources algorithms that define how the application system uses the resources We assume that workload characteristics are given to us a priori before the execution begins © Cheol-Hoon Lee SSLab, CNU, 2015 Ch.3 A Reference Model of Real-Time Systems

Processors and Resources We divide all the system resources into processors and resources Processors are often called servers and active resources computers, transmission links, disks, and database server denoted by Every job must have one or more processors to execute Speed is the main attribute of a processor Resources mean passive resources memory, sequence numbers, mutexes, and database locks A job may need some resources in addition to the processor (ex. semaphore) Each resource may have one or more units © Cheol-Hoon Lee SSLab, CNU, 2015 Ch.3 A Reference Model of Real-Time Systems

Temporal Parameters of Real-Time Workload The workload on processors consists of jobs job : a unit of work to be allocated processor time and other resources task : a set of related jobs that execute to support a function of the system Each job is characterized by temporal parameter : timing constraints and behavior functional parameter : intrinsic properties of the job resource parameter : resource requirement interconnection parameter : dependency with other jobs Temporal parameter of a job release time (or arrival time) release-time jitter absolute deadline relative deadline execution time or © Cheol-Hoon Lee SSLab, CNU, 2015 Ch.3 A Reference Model of Real-Time Systems

Ch.3 A Reference Model of Real-Time Systems Periodic Task Model A period task is defined by the following task parameters period : the minimum length of the interval times of consecutive jobs (i.e., interrelease interval) execution time : the maximum execution time of all the jobs in phase : the release time of the first job in Other definitions hyperperiod : the least common multiple of for utilization of task : total utilization Deadline is usually assumed to be equal to its period (i.e., ) However, in general, it can be shorter than © Cheol-Hoon Lee SSLab, CNU, 2015 Ch.3 A Reference Model of Real-Time Systems

Aperiodic and Sporadic Tasks A real-time system is invariably required to respond to external events, and to respond, it executes aperiodic or sporadic jobs whose release time is not known a priori. A task is aperiodic if the jobs in it have either soft deadline or no deadlines We want to optimize the responsiveness of the system for the aperiodic jobs, but never at the expense of hard real-time tasks whose deadline must be met at all time Tasks containing jobs that are released at random time instants and have hard deadlines are sporadic tasks Our primary concern is to ensure that their deadlines are always met; minimizing their response times is of secondary importance © Cheol-Hoon Lee SSLab, CNU, 2015 Ch.3 A Reference Model of Real-Time Systems

Precedence Constraints The jobs are said to have precedence constraints if they are constrained to execute in some order precedence relation A job is a predecessor of another job (and is a successor of ) if cannot begin execution until the execution of completes is an immediate predecessor of (and is an immediate successor of ) if and there is no other job such that A job with predecessors is ready for execution when the time is at or later its release time and all of its predecessors are completed precedence graph G=(J,<) A directed graph to represent the precedence constraints among jobs each vertex : a job in J There is an directed edge from to when is an immediate predecessor of task graph G=(V,E) An extended precedence graph to describe the application system (see Fig 3-1) © Cheol-Hoon Lee SSLab, CNU, 2015 Ch.3 A Reference Model of Real-Time Systems

Data and Other Dependencies Data dependency In many real-time systems, jobs communicate via shared data Each producer places the data generated by it in a shared address space to be used by the consumer at any time (ex. navigation job and flight management job) A parameter of an data dependency edge is the volume of the data Temporal dependency Some jobs may be constrained to complete within a certain amount of time interval relative to one another temporal distance: the difference in the completion times of two jobs Jobs are said to have a temporal distance constraint if their temporal distance must be no more than some finite value (ex. display of video frames and the accompanying audio) © Cheol-Hoon Lee SSLab, CNU, 2015 Ch.3 A Reference Model of Real-Time Systems

Data and Other Dependencies - continued AND/OR precedence constraints AND job : a job with more than one immediate predecessors must wait until all its predecessors have been completed before its execution OR job : one which can begin execution if one or some its immediate predecessors has been completed Conditional branches Conditional execution can be represented by branch and join jobs (cf. OR constraints versus AND constraints) Pipeline relationship There is an edge from to if the output of is piped into and the execution of can proceed as long as there are data for it to process © Cheol-Hoon Lee SSLab, CNU, 2015 Ch.3 A Reference Model of Real-Time Systems

Ch.3 A Reference Model of Real-Time Systems Example of Task Graph (0, 7] (2, 9] (4, 11] (6,13] (8,15] (2, 5] (5, 8] (8, 11] (11, 14] (14, 17] (0, 5] (4, 8] (5, 20] Conditional block branch join (0, 6] J (2, 10] 2/3 1/2 Fig. 3-1 Example of task graphs © Cheol-Hoon Lee SSLab, CNU, 2015 Ch.3 A Reference Model of Real-Time Systems

Functional Parameters Preemptivity of jobs A job is preemptable if its execution time can be suspended at any time to allow the execution of other jobs and, later on, can be resumed from the point of suspension A job may be preemptable everywhere except for a small portion which is constrained to be nonpreemptable (ex.) interrupt handling job (why?) Context-switch time is the preemption overhead Criticality of jobs The importance (or criticality) of a job is a positive number that indicates how critical the job is with respect to other jobs During an overload when it is not possible to schedule all the jobs to meet their deadlines, it may make sense to sacrifice the less critical jobs so that the more critical jobs can meet their deadlines (ex.) flight control job > navigation job > cruise control job weighted average response time or weighted average tardiness © Cheol-Hoon Lee SSLab, CNU, 2015 Ch.3 A Reference Model of Real-Time Systems

Functional Parameters - continued Optional executions Some jobs or portions of jobs may be indicated as optional If an optional job or an optional portion of a job completes late or is not executed at all, the system performance may degrade, but nevertheless function satisfactorily Mandatory jobs (or portions of jobs) must be executed to completion During a transient overload when it is not possible to complete all the jobs in time, we may choose to discard optional jobs so that the mandatory jobs can complete in time (ex.) In a collision avoidance system, we may consider the job that computes the correct evasive action and informs the operator optional While the job that generates a warning and displays the course of the object about to collide with it in time mandatory © Cheol-Hoon Lee SSLab, CNU, 2015 Ch.3 A Reference Model of Real-Time Systems

Functional Parameters - continued Laxity type and laxity function The laxity type indicates whether its timing constraints are soft or hard It can be sometimes supplemented by a usefulness function It gives the usefulness of the result produced by the job as a function of its tardiness Usefulness functions can guide the choice and implementation of scheduling strategies Value 1, 2. hard real-time jobs 2. “better never than late” 3. point-of-sales transaction 4. stock price update transaction 3 4 tardiness 1 2 Fig. 3-3 Example of usefulness functions © Cheol-Hoon Lee SSLab, CNU, 2015 Ch.3 A Reference Model of Real-Time Systems

Resource Parameters of Jobs A resource parameter is preemptivity If jobs can use every unit of a resource in an interleaved fashion, the resource is preemptable A resource is nonpreemptable, if each unit of the resource is constrained to be used serially In other words, once a unit of a nonpreemptable resource is allocated to a job, other jobs needing the unit must wait until the job completes its use (ex.) the lock on a data object in a database Resource graph A resource graph describes the configuration of the resources vertex : processor or (passive) resource is-a-part-edge A memory is a part of a computer, and so is a monitor accessibility edge If there is a connection between two CPUs in the two computers, then each CPU is accessible from the other computer A parameter is the cost for sending a unit of data © Cheol-Hoon Lee SSLab, CNU, 2015 Ch.3 A Reference Model of Real-Time Systems

Model of Real-Time Systems Three elements of the model of real-time systems The application system is represented by a task graph It gives the processor time and resource requirements of jobs, the timing constraints of each job, and the dependencies of jobs A resource graph describes the configuration of the resources It describes the amounts of the resources available to execute the application system, the attributes of the resources, and the rules governing their usage Between them are the scheduling and resource access-control algorithms used by the operating system © Cheol-Hoon Lee SSLab, CNU, 2015 Ch.3 A Reference Model of Real-Time Systems

Model of Real-Time Systems - continued scheduling and resource access-control processors resources Fig. 3-4 Model of real-time systems © Cheol-Hoon Lee SSLab, CNU, 2015 Ch.3 A Reference Model of Real-Time Systems

Scheduler and Schedules The scheduler assigns processors to jobs A schedule means an assignment of all the jobs in the system on the available processors produced by the scheduler We assume that the scheduler produces only valid schedules: Every processor is assigned to at most one job at any time Every job is assigned at most one processor at any time No job is scheduled before its release time Depending on the scheduling algorithm(s) used, the total amount of processor time assigned to every job is equal to its maximum or actual execution time All the precedence and resource usage constraints are satisfied An implicit assumption is that jobs do not run in parallel on more than one processor to speed up their execution © Cheol-Hoon Lee SSLab, CNU, 2015 Ch.3 A Reference Model of Real-Time Systems

Feasibility and Optimality A valid schedule is feasible if every job completes by its deadline We say a set of jobs is schedulable according to a scheduling algorithm if when using the algorithm the scheduler always produces a feasible schedule A hard real-time scheduling algorithm is optimal if the algorithm always produces a feasible schedule if there are feasible schedules Conversely, if an optimal algorithm fails to find a feasible schedule, the given set of jobs cannot feasibly be scheduled by any algorithm © Cheol-Hoon Lee SSLab, CNU, 2015 Ch.3 A Reference Model of Real-Time Systems

Ch.3 A Reference Model of Real-Time Systems Performance Measures Lateness vs. tardiness The lateness is the difference between completion time and deadline It can be a negative value (What about tardiness?) (ex.) packet transmission in a packet-switched network Makespan What if when all the jobs have the same release time and deadline? If the makespan is less than or equal to the length of their feasible interval, the jobs can meet their deadline Average response time The most commonly used performance measure for soft real-time jobs In a system that has a mixture of jobs with hard and soft deadlines, the objective of scheduling is typically to minimize the average response time of soft deadline jobs while ensuring all hard deadline jobs complete in time Miss rate, loss rate, and invalid rate miss rate: percentage of jobs that are executed but completed too late loss rate: percentage of jobs that are not executed at all invalid rate: miss rate + loss rate © Cheol-Hoon Lee SSLab, CNU, 2015 Ch.3 A Reference Model of Real-Time Systems