Real-Time Systems. Example: scheduling using EDF. Feasibility analysis for EDF. Example: scheduling using EDF

Similar documents
10/04/2018. Definitions. The scheduling problem. Feasibility vs. schedulability. Complexity. Why do we care about complexity? P R

FORCED VIBRATION of MDOF SYSTEMS

termination suspend (swap-out) executing CPU Scheduling phases (non real-time) pre-emption or cycle done unblock ready, suspended

14. Poisson Processes

The Poisson Process Properties of the Poisson Process

8. Queueing systems lect08.ppt S Introduction to Teletraffic Theory - Fall

Redundancy System Fault Sampling Under Imperfect Maintenance

Cyclone. Anti-cyclone

Determination of Antoine Equation Parameters. December 4, 2012 PreFEED Corporation Yoshio Kumagae. Introduction

Midterm Exam. Tuesday, September hour, 15 minutes

The ray paths and travel times for multiple layers can be computed using ray-tracing, as demonstrated in Lab 3.

QR factorization. Let P 1, P 2, P n-1, be matrices such that Pn 1Pn 2... PPA

Quantum Mechanics II Lecture 11 Time-dependent perturbation theory. Time-dependent perturbation theory (degenerate or non-degenerate starting state)

Continuous Time Markov Chains

Least Squares Fitting (LSQF) with a complicated function Theexampleswehavelookedatsofarhavebeenlinearintheparameters

The Mean Residual Lifetime of (n k + 1)-out-of-n Systems in Discrete Setting

For the plane motion of a rigid body, an additional equation is needed to specify the state of rotation of the body.

Chapter 8. Simple Linear Regression

(1) Cov(, ) E[( E( ))( E( ))]

The Linear Regression Of Weighted Segments

FALL HOMEWORK NO. 6 - SOLUTION Problem 1.: Use the Storage-Indication Method to route the Input hydrograph tabulated below.

Optimal Eye Movement Strategies in Visual Search (Supplement)

Partial Molar Properties of solutions

Queuing Theory: Memory Buffer Limits on Superscalar Processing

-distributed random variables consisting of n samples each. Determine the asymptotic confidence intervals for

Exam Supply Chain Management January 17, 2008

Key words: Fractional difference equation, oscillatory solutions,

Lecture 3 Topic 2: Distributions, hypothesis testing, and sample size determination

International Journal Of Engineering And Computer Science ISSN: Volume 5 Issue 12 Dec. 2016, Page No.

Model for Optimal Management of the Spare Parts Stock at an Irregular Distribution of Spare Parts

SYRIAN SEISMIC CODE :

Solution. The straightforward approach is surprisingly difficult because one has to be careful about the limits.

θ = θ Π Π Parametric counting process models θ θ θ Log-likelihood: Consider counting processes: Score functions:

Linear Regression Linear Regression with Shrinkage

Fundamentals of Speech Recognition Suggested Project The Hidden Markov Model

Stabilization of LTI Switched Systems with Input Time Delay. Engineering Letters, 14:2, EL_14_2_14 (Advance online publication: 16 May 2007) Lin Lin

Big O Notation for Time Complexity of Algorithms

4. THE DENSITY MATRIX

Solution set Stat 471/Spring 06. Homework 2

ODEs II, Supplement to Lectures 6 & 7: The Jordan Normal Form: Solving Autonomous, Homogeneous Linear Systems. April 2, 2003

Quantitative Portfolio Theory & Performance Analysis

Available online Journal of Scientific and Engineering Research, 2014, 1(1): Research Article

UNIVERSITY OF OSLO DEPARTMENT OF ECONOMICS

Competitive Facility Location Problem with Demands Depending on the Facilities

IMPROVED PORTFOLIO OPTIMIZATION MODEL WITH TRANSACTION COST AND MINIMAL TRANSACTION LOTS

Éditeur Inria, Domaine de Voluceau, Rocquencourt, BP 105 LE CHESNAY Cedex (France) ISSN

For combinatorial problems we might need to generate all permutations, combinations, or subsets of a set.

Solution of Impulsive Differential Equations with Boundary Conditions in Terms of Integral Equations

Solving fuzzy linear programming problems with piecewise linear membership functions by the determination of a crisp maximizing decision

USING INPUT PROCESS INDICATORS FOR DYNAMIC DECISION MAKING

Fully Fuzzy Linear Systems Solving Using MOLP

Written Exercise Sheet 5

Introduction to local (nonparametric) density estimation. methods

Application of the stochastic self-training procedure for the modelling of extreme floods

Cyclically Interval Total Colorings of Cycles and Middle Graphs of Cycles

Interval Estimation. Consider a random variable X with a mean of X. Let X be distributed as X X

Least squares and motion. Nuno Vasconcelos ECE Department, UCSD

Mathematical Formulation

The Selection Problem - Variable Size Decrease/Conquer (Practice with algorithm analysis)

Periodic Resource Reallocation in Two-Echelon Repairable Item Inventory Systems

AML710 CAD LECTURE 12 CUBIC SPLINE CURVES. Cubic Splines Matrix formulation Normalised cubic splines Alternate end conditions Parabolic blending

The Bernstein Operational Matrix of Integration

The textbook expresses the stock price as the present discounted value of the dividend paid and the price of the stock next period.

The textbook expresses the stock price as the present discounted value of the dividend paid and the price of the stock next period.

Complete Identification of Isotropic Configurations of a Caster Wheeled Mobile Robot with Nonredundant/Redundant Actuation

NUMERICAL EVALUATION of DYNAMIC RESPONSE

Fault Tolerant Computing. Fault Tolerant Computing CS 530 Probabilistic methods: overview

Economics 8723 Macroeconomic Theory Problem Set 2 Professor Sanjay Chugh Spring 2017

VARIATIONAL ITERATION METHOD FOR DELAY DIFFERENTIAL-ALGEBRAIC EQUATIONS. Hunan , China,

PTAS for Bin-Packing

Functions of Random Variables

Pricing of CDO s Based on the Multivariate Wang Transform*

Reliability Equivalence of a Parallel System with Non-Identical Components

THE PUBLISHING HOUSE PROCEEDINGS OF THE ROMANIAN ACADEMY, Series A, OF THE ROMANIAN ACADEMY Volume 10, Number 2/2009, pp

Chapter 3: Maximum-Likelihood & Bayesian Parameter Estimation (part 1)

SPECIAL CONSIDERATIONS FOR VOLUMETRIC Z-TEST FOR PROPORTIONS

Extremal graph theory II: K t and K t,t

The algebraic immunity of a class of correlation immune H Boolean functions

Brownian Motion and Stochastic Calculus. Brownian Motion and Stochastic Calculus

CS 268: Packet Scheduling

As evident from the full-sample-model, we continue to assume that individual errors are identically and

Bianchi Type II Stiff Fluid Tilted Cosmological Model in General Relativity

General Complex Fuzzy Transformation Semigroups in Automata

NOTE ON SIMPLE AND LOGARITHMIC RETURN

Common MidPoint (CMP) Records and Stacking

Handling on-line changes. Handling on-line changes. Handling overload conditions. Handling on-line changes

Third handout: On the Gini Index

Chapter 8: Statistical Analysis of Simulated Data

Laboratory I.10 It All Adds Up

Density estimation III.

A Constitutive Model for Multi-Line Simulation of Granular Material Behavior Using Multi-Plane Pattern

Dynamic Voltage and Frequency Scaling Under a Precise Energy Model Considering Variable and Fixed Components of the System Power Dissipation

Real-time Classification of Large Data Sets using Binary Knapsack

Solving Constrained Flow-Shop Scheduling. Problems with Three Machines

Fresnel Equations cont.

Thus, lmted mache avalablty s commo practce. Kowledge about mache avalabltes mght be complete or complete. I a ole settg mache avalabltes are ot kow a

A tighter lower bound on the circuit size of the hardest Boolean functions

Outline. Computer Networks: Theory, Modeling, and Analysis. Delay Models. Queuing Theory Framework. Delay Models. Little s Theorem

Exercise 3 Stochastic Models of Manufacturing Systems 4T400, 6 May

Comparison of the Bayesian and Maximum Likelihood Estimation for Weibull Distribution

Transcription:

EDA/DIT6 Real-Tme Sysems, Chalmers/GU, 0/0 ecure # Updaed February, 0 Real-Tme Sysems Specfcao Problem: Assume a sysem wh asks accordg o he fgure below The mg properes of he asks are gve he able Ivesgae he schedulably of he asks whe EDF s used (Noe ha D < T for all asks) Implemeao Verfcao Dyamc schedulg -- Earles-deadle-frs schedulg Processor-demad aalyss Task C D T 8 Feasbly aalyss for EDF Smulae a execuo of he asks: 0 6 8 The asks are o schedulable eve hough U = + + 8 = 7 8 = 0875 < msses s deadle! Wha aalyss mehods are suable for geeral EDF: Ulzao-based aalyss? No suable! No geeral eough or exac eough Does o work well for he case of D < T Respose-me aalyss? No suable! Aalyss much more complex ha for DM Crcal sa does o ecessarly occur whe all asks arrve a he same me for he frs me Isead, respose me of a ask s maxmzed a some scearo where all oher asks arrve a he same me; he wors such scearo has o be defed for each ask before he respose me of ha ask ca be calculaed

EDA/DIT6 Real-Tme Sysems, Chalmers/GU, 0/0 ecure # Updaed February, 0 Processor-demad aalyss Processor demad: The processor demad for a ask a gve me erval [ 0, ] s he amou of processor me ha he ask eeds he erval order o mee he deadles ha fall wh he erval e N represe he umber of saces of ha mus complee execuo before The oal processor demad up o s (0, ) = = N C Processor-demad aalyss Processor demad: N We ca calculae by coug how may mes ask has arrved durg he erval 0, D We ca gore saces of he ask ha arrved durg he erval D, sce D > for hese saces 0 Isace wh D > N = N = Processor-demad aalyss Processor-demad aalyss: N We ca express as N = D T + The oal processor demad s hus (0, ) = = D T + C Exac feasbly es for EDF (Suffce ad ecessary codo) A suffce ad ecessary codo for earlesdeadle-frs schedulg, for whch, s D T : (0, ) where (0, ) s he oal processor demad [ 0, ] The processor-demad aalyss ad assocaed feasbly es was preseed by S Baruah, Roser ad R Howell 990

EDA/DIT6 Real-Tme Sysems, Chalmers/GU, 0/0 ecure # Updaed February, 0 Exac feasbly es for EDF (Suffce ad ecessary codo) How large erval mus be examed? arges erval s CM of he ask s perods Ths erval ca be furher shoreed (see course book) How may me pos mus be examed? Oly absolue deadles eed o be examed The se of deadles ca be furher reduced (see course book) The es ca cosequely be mproved as follows: K = K : (0, ) { D k D k = kt + D, D k CM{ T,,T },, k 0} Problem: Assume a sysem wh asks accordg o he fgure below The mg properes of he asks are gve he able a) Deerme, by aalyzg he processor demad, wheher he asks are schedulable or o usg EDF b) Deerme, by usg smulao, wheher he asks are schedulable or o usg EDF We solve hs o he blackboard! Task C D T 7 8 6 Exeded processor-demad aalyss Deadle verso The es ca be exeded o hadle: Blockg Sar-me varaos ( release jer ) Tme offses H ormal execuo crcal rego H blocked H ad share resource R H msses s deadle I hs course, we oly sudy blockg M

EDA/DIT6 Real-Tme Sysems, Chalmers/GU, 0/0 ecure # Updaed February, 0 Sack Resource Polcy (SRP) The asks are assged preempo levels, he properes of whch are: The preempo level of ask s deoed π Task s o allowed o preemp aoher ask j uless π If has hgher prory ha j ad arrves laer, he mus have a hgher preempo level ha Noe: - The preempo levels are sac values, eve hough he asks prores may be dyamc - For EDF schedulg, suable levels ca be derved f asks wh shorer relave deadles ge hgher preempo levels, ha s: π D < D j j Sack Resource Polcy (SRP) Deadle verso ca be reduced wh resource celgs: Each shared resource s assged a celg ha s always equal o he maxmum preempo level amog all asks ha may be blocked whe requesg he resource The proocol keeps a sysem-wde celg ha s equal o he maxmum of he curre celgs of all resources A ask wh he earles deadle s allowed o preemp oly f s preempo level s hgher ha he sysem-wde celg Noe: The orgal prory of he ask s o chaged a ru-me The resource celg s a dyamc value calculaed a ru-me as a fuco of curre resource avalably Oherwse, he behavor of he SRP proocol s very smlar o he ICPP, ad SRP also exhbs decal properes regardg maxmum blockg me ad freedom from deadlock Sack Resource Polcy (SRP) Exeded processor-demad aalyss H M preempo level (H) > preempo level (M) > preempo level () ormal execuo H ad share resource R crcal rego sysem-wde celg s se o R s resource celg (= H s preempo level) H blocked sysem-wde celg s resored Blockg ca be accoued for he followg way: B Blockg facor represes he legh of crcal / opreempve regos ha are execued by asks wh lower preempo levels ha Tasks are dexed he order of creasg preempo levels, ha s: π < j K, [, ]: (0, ) Aga: oe he smlary o he behavor of ICPP D = k T k + C k + k = D T + B

EDA/DIT6 Real-Tme Sysems, Chalmers/GU, 0/0 ecure # Updaed February, 0 Exeded processor-demad aalyss Deermg he blockg facor for ask : Deerme he wors-case resource celg for each crcal rego, ha s, assume he ru-me suao where he correspodg resource s uavalable Idefy he asks ha have a preempo level lower ha ad ha calls crcal regos wh a wors-case resource celg equal o or hgher ha he preempo level of Cosder he mes ha hese asks lock he acual crcal regos The loges of hose mes cosues he blockg facor B Problem: Assume a sysem wh asks accordg o he fgure below The mg properes of he asks are gve he able Three resources R, R ad R have hree, oe, ad hree us avalable, respecvely The parameers H R, H R ad H R represe he loges me a ask may use he correspodg resource The parameers μ R, μ R ad μ R represe he umber of us a ask requess from he correspodg resource Task Task C D D T 6 0 50 7 7 50 R R R 0 5 50 H R H R - H R μ R μ R μ R - Feasbly ess Problem: (co d) Task frs requess R ad he, whle usg R, requess R Task frs requess R ad he, whle usg R, requess R; he, afer releasg he wo resources, requess R Task frs requess R ad he, whle usg R, requess R; he, afer releasg he wo resources, requess R Exame he schedulably of he asks whe he SRP (Sack Resource Polcy) proocol s used a) Derve he celgs (dyamc ad wors-case) of he resources b) Derve he blockg facors for he asks c) Show wheher he asks are schedulable or o We solve hs o he blackboard! Sac prory (RM/DM) Dyamc prory (EDF) U ( / ) U Summary D = T D T : R = C + : = j hp() D T R T j C j D + C 5