From UML/MARTE Models of Multiprocessor Real-time Embedded

Systems to Early Schedulability Analysis based on SimSo Tool

Amina Magdich, Yessine Hadj Kacem, Adel Mahfoudhi and Mohamed Abid

CES laboratory, National School of Engineers of Sfax (ENIS), Sfax, Tunisia

Keywords:

MARTE, MDE, Model Transformation, SimSo, Semi-partitioned Scheduling, Global Scheduling.

Abstract:

The increasing complexity of Real-Time Embedded Systems (RTES) should be met with equally sophisticated

design methods. The recent Uniﬁed Modeling Language (UML) proﬁle for Modeling and Analysis of Real-

Time Embedded systems (MARTE) is well adapted for systems modeling. However along with the variety

of schedulability analysis tools, bridging the gap between design models and meta-models of the documented

scheduling analysis tools becomes an important issue.

In this paper, we discuss a Model To Text (M2T) transformation for enabling the derivation of schedulability

analysis models from UML/MARTE models. The generated model for schedulability analysis represents an

input for an analysis tool. As a proof of concepts, we present the implemented code and experimental results.

1 INTRODUCTION

The spread of technology and the industry re-

quirements have pushed designers to switch from

simple monoprocessor architectures to more complex

parallel multiprocessor architectures. Considering

multiprocessor systems leads to an increasing trend of

RTES design, which requires rigorous methodologies

to reduce the designer’s effort and avoid systems

failures. A prominent effort has been focused on the

use of Model Driven Engineering (MDE) (Schmidt,

2006) and high-level modeling languages such as

UML/MARTE proﬁle (OMG, 2008) to automate the

design ﬂows and raise the abstraction level.

On the other hand, designers are interested in verify-

ing the temporal correctness of their studied systems

at early design stages to be reassured that no deadline

may be missed. In this context, the schedulability

analysis is used to validate the temporal behavior of

systems scheduled using monoprocessor or multipro-

cessor scheduling approaches.

Regarding multiprocessor scheduling, three ap-

proaches are available in the literature; the partitioned

scheduling approach, the global scheduling approach

and the semi-partitioned one (Dorin et al., 2010).

The partitioned scheduling approach consists on

statically assigning each task to be executed on only

one processor. Using this strategy comes to using

monoprocessor scheduling approach, since each

task may be allocated to only one processor. In a

partitioned scheduling context, tasks are not allowed

to migrate inter-processors.

While adopting a global scheduling approach, tasks

are dynamically allocated to processors and they

are allowed to migrate inter-processors improving

then Central Processing Units (CPUs) occupation.

While using this approach, a full migration of tasks is

allowed. Consequently, an attention must be given to

the cost of preemption and context switching as well

as the number of cache misses due to the transferring

of tasks from one computing resource to another one.

Under the semi-partitioned scheduling approach,

most tasks are assigned to be executed on speci-

ﬁed processors like in the partitioned scheduling

approach. Nevertheless, tasks that may not be

assigned to a single processor are allowed to migrate

inter-processors. This approach enables a restricted

task migration to maximize CPU occupation and

reduce context switching costs.

Regardless the used scheduling approach, the

schedulability analysis has always been an important

issue that has been widely studied during the last

years. Nevertheless, there are still many open issues

regarding this context. In fact, due to the variety of

schedulability analysis tools coupled with the ever

growing complexity of RTES, there is still a need

to automate the early schedulability analysis step to

reduce designers’ effort. Researchers’ attention has

been then focused on the transformation of systems

models into analysis tools meta-models to analyze

202

Magdich, A., Kacem, Y., Mahfoudhi, A. and Abid, M.

From UML/MARTE Models of Multiprocessor Real-time Embedded Systems to Ear ly Schedulability Analysis based on SimSo Tool.

DOI: 10.5220/0005982902020209

In Proceedings of the 11th International Joint Conference on Software Technologies (ICSOFT 2016) - Volume 1: ICSOFT-EA, pages 202-209

ISBN: 978-989-758-194-6

systems, which are scheduled using monoprocessor

and multiprocessor scheduling approaches.

Nevertheless, considering the multiprocessor

scheduling, the documented automatic schedulability

analysis approaches have addressed only the parti-

tioned and global scheduling approaches.

In this context, we propose an automatic schedula-

bility analysis for systems that are scheduled using

semi-partitioned and global scheduling approaches.

MDE is used in this context to raise the abstraction

level and automate the schedulability analysis step.

The remainder of this paper starts with motivation

and related works in section 2. Section 3 gives an

overview of high-level methodologies and analysis

tools for early schedulability analysis. In section 4,

the proposed process for automatic schedulability

analysis at early design stages is highlighted. To

validate our proposal, a case study is performed in

section 5. Finally, conclusions and future works are

given in section 6.

2 MOTIVATION AND RELATED

WORK

Checking if RTES meet their timing requirements at

early design stages is extremely important to avoid

systems failures. In this context, a wide number of re-

search works have been proposed (Zhang and Burns,

2009)(Abdeddaïm et al., 2014)(Lee and Shin, 2013).

On the other hand, along with the variety of schedula-

bility analysis tools, the schedulability analysis is still

considered as an important issue that needs to be man-

aged properly at a high-level of abstraction to reduce

designers’ effort.

In fact, each schedulability analysis tool is based on

an input meta-model that encloses all the required in-

formation for schedulability analysis step.

An input meta-model may be deﬁned using UML,

SysML, MARTE, XML, etc.

Since input meta-models differ from one analysis tool

to another, some research works have been focused

on the building of speciﬁc models that represent in-

put for the adopted analysis tools such as in (Jensen,

2009). In this proposal, authors have used MARTE

to build a speciﬁc model that represents an input for

MAST tool. This methodology has been used to sup-

port only monoprocessor systems. Moreover, the con-

struction of input models for analysis tools has been

done manually. With this regard, other research works

have been focused on the automatic built of analysis

models through a transformation of systems models

to schedulability analysis meta-models.

This model transformation offers automatic schedu-

lability analysis step and fosters the independence of

the design ﬂow towards the used analysis tools.

In this context, a Model To Model (M2M) transfor-

mation has been performed in (Hagner and Huhn,

2008) to establish an early schedulability analysis of

systems using SymTA/S tool (Henia et al., 2005).

MARTE proﬁle has been used in this context to an-

notate models with timing requirements. Activity

diagrams annotated through Schedulability Analysis

Modeling (SAM) have been transformed to SymTA/S

tool meta-model. In the same context, OPTIMUM

methodology has been provided in (Mraidha et al.,

2011) for early schedulability analysis of systems

modeled using MARTE sub-proﬁles mainly Generic

Resource Modeling (GRM) and SAM.

The models transformation that has been proposed

in this methodology enables generating concurrency

models that may be analyzed and validated using

COTS schedulability analysis tools. OPTIMUM has

been tested in the case of monoprocessor systems.

In (Medina and Cuesta, 2011), the MARTE proﬁle

has been used to model temporal requirements of

multiprocessor systems. Built models are then trans-

formed to cheddar tool meta-model to establish an

early schedulability analysis for multiprocessors sys-

tems while considering partitioned scheduling.

In the same context, a model to model transformation

from an activity diagram to Petri Net tool has been

proposed in (HadjKacem et al., 2012) to establish an

early schedulability analysis of RTES. MARTE/SAM

sub-proﬁle has been used to annotate the dynamic

view. The multiprocessor scheduling regarding the

partitioned scheduling approach has been supported.

In (Naija et al., 2015), an early schedulability anal-

ysis for real-time systems has been proposed while

using MDE concepts. To perform this step, a model

to model transformation has been performed to trans-

form an activity diagram annotated with SAM to Petri

Nets tool meta-model. Only the partitioned schedul-

ing has been supported in this proposal.

In (Rubini et al., 2013), an early schedulability anal-

ysis from AADL models is performed while using

cheddar tool. In this proposal, cheddar has been ex-

tended to support global scheduling, but it stills does

not support semi-partitioned scheduling.

In previous cited research works for automatic

schedulability analysis at early design stages, mono-

processor and multiprocessor scheduling have been

supported. Nevertheless, considering the multipro-

cessor scheduling, only the partitioned or global

scheduling approaches have been addressed. No

attention has been given to the semi-partitioned

scheduling approach. With this regard, we propose in

this paper an early schedulability analysis for semi-

From UML/MARTE Models of Multiprocessor Real-time Embedded Systems to Early Schedulability Analysis based on SimSo Tool

203

partitioned and global scheduling. Our proposal is

based on the use of MDE concepts and UML/MARTE

proﬁle for a high-level automatic schedulability anal-

ysis. Our main goal is to establish models transforma-

tion in order to reduce the gap between systems mod-

els and schedulability analysis tools meta-models.

3 HIGH-LEVEL

METHODOLOGIES AND

SCHEDULABILITY ANALYSIS

TOOLS FOR AUTOMATIC

SCHEDULABILITY ANALYSIS

OF MULTIPROCESSOR

SYSTEMS

While dealing with complex systems, the use of high-

level languages and techniques reduces designers’ ef-

fort and overcomes the design challenges.

In this section, we give an overview of the used lan-

guages, techniques and tools to establish a high-level

schedulability analysis step.

3.1 MDE

The Model Driven Engineering (Schmidt, 2006) is a

software development methodology considered to be

an effective solution that simpliﬁes the design pro-

cess since it focuses on the abstract representation

of domains rather than on computing concepts such

as algorithmic concepts. The promoted idea of the

MDE paradigm is to use models at different level of

abstraction while designing systems to raise the ab-

straction level of systems speciﬁcation and increase

the automation of their development. MDE is based

on three main concepts which are meta-model, model

and model transformation.

A model is often speciﬁed using Domain Speciﬁc

Language (DSL) that can be graphical or textual such

as UML, MARTE, etc. In fact, different transforma-

tion techniques are available in the literature such as

M2T and M2M.

The M2T type, which represents a mapping from

a model to a text, is based on existing parsers

(such as XML/XSLT) that are based on pro-

gramming languages (JAVA) or mapping templates

(JET/ACCELEO). Regarding the M2M technique, it

uses mapping languages (ATL or Kermeta) to trans-

late a meta-model to another meta-model while adopt-

ing a syntactic and semantic analysis.

In our proposal, we adopt the M2T transformation

to transform the system model into a schedulability

analysis meta-model and then establish an automatic

schedulability analysis at early design stages.

In this context, we have implemented an ACCELEO

template (ECLIPSE, 2008) to perform this transfor-

mation.

3.2 MARTE

MARTE is a proﬁle adopted by the Object Manage-

ment Group (OMG) to replace the UML Proﬁle for

Schedulability, Performance and Time (SPT) (OMG,

2002). This proﬁle deﬁnes foundations to support

speciﬁcation, design and veriﬁcation of RTES. It pro-

vides a common way of modeling both hardware and

software aspects. MARTE is considered as one of the

most commonly used high-level language for com-

plex systems modeling. Considerably, it models prop-

erly temporal requirements of RTES.

MARTE encloses different sub-proﬁles providing a

big set of stereotypes and attributes to annotate mod-

els with data, which are required to establish speciﬁc

analysis. Among these sub-proﬁles, we cite Generic

Resource Modeling (GRM), Software Resource Mod-

eling (SRM), Hardware Resource Modeling (HRM),

Generic Quantitative Analysis Modeling (GQAM),

Performance Analysis Modeling (PAM), SAM, etc.

What is worthwhile to note is that originally

MARTE supported only monoprocessor and parti-

tioned scheduling.

In this context, extensions of MARTE have been doc-

umented to support both semi-partitioned and global

scheduling (Magdich et al., 2012)(Magdich et al.,

2013a)(Magdich et al., 2013b). These extensions are

used in this paper to model the studied system.

3.3 Schedulability Analysis Tools for

Multiprocessor Scheduling

To validate temporal requirements of critical applica-

tions, various tools for scheduling analysis have been

documented in the literature such as MAST (Gonza-

lez Harbour et al., 2001), Cheddar (Singhoff et al.,

2004), (Rubini et al., 2013), RealtssMP (Ramirez

et al., 2012), STORM Simulation TOol for Real-time

Multiprocessor scheduling (Urunuela et al., 2010),

Colored Petri Nets (CPN) tools (cpn, 2015), etc.

The cited scheduling analysis tools may be used to

analyze systems which are scheduled using monopro-

cessor or multiprocessor scheduling approaches.

Considering multiprocessor scheduling approaches,

only the partitioned and global scheduling approaches

are supported by these tools. On the other hand, these

tools are designed to validate, test and analyze sys-

tems without handling with direct overheads such as

ICSOFT-EA 2016 - 11th International Conference on Software Engineering and Applications

204

Figure 1: Early schedulability analysis methodology based on MDE and Simso tool.

scheduling overheads and context switching.

To deal with this issue, SimSo tool for Simula-

tion of Multiprocessor Scheduling with Overheads

(Chéramy et al., 2014) has been proposed. This open

source tool is used in our proposal since it supports

both semi-partitioned and global scheduling. The in-

put of the considered tool is an XML ﬁle that encloses

all the needed criteria for temporal veriﬁcation.

Consequently, the challenge addressed in this paper

is the transformation of a MARTE model supporting

semi-partitioned or global scheduling to SimSo meta-

model for early and automatic schedulability analysis.

4 PROPOSED METHODOLOGY

FOR AUTOMATIC

SCHEDULABILITY ANALYSIS

AT AN EARLY STAGE

The proposed methodology enables automatic

schedulability analysis of RTES at early design

stages while supporting both semi-partitioned and

global scheduling approaches. Figure 1 shows the

different steps adopted by our methodology to check

the temporal correctness of RTES.

To validate temporal correctness of RTES, SimSo

tool that accepts only XML ﬁle as input is used in our

proposal.

Given a system model annotated through MARTE

proﬁle and mainly using GRM and GQAM (Step1),

a M2T transformation must be performed to translate

the system properties from the MARTE model to the

meta-model of SimSo tool. In this context, we have

implemented an ACCELEO template (Figure 2) to

support this transformation (Step2).

This template implements transformation of MARTE

concepts to SimSo meta-model concepts.

The execution of the implemented template allows

the generation of an XML ﬁle, which contains the

system properties (Step3).

This ﬁle will be entered to the SimSo tool for

schedulability analysis check (Step4). In case of

non schedulability, a feed-back has to be done for

MARTE model rectiﬁcation. Table1 shows some of

transformation concepts from MARTE to SimSo.

Table 1: MARTE to SimSo transformation concepts.

MARTE SimSo

Processor

stereotypes

«HwProcessor» processor

and

annotations

«HwComputing

Resource»

«name» name

«speedFactor» speed

Tasks

stereotypes

and

annotations

«SwSchedulable

Resource»

task

«isPreemptable» preemptible

«type» task_type

«periodElements» period

Scheduler

stereotypes

«GaExecHost» sched

and

annotations

«Scheduler»

«otherSchedPolicy» class

For example, a class annotated through the stereotype

«HwProcessor» or «HwComputingResource» (or

both of them) models a processor. It is transformed

to an element named processor.

«speedFactor» is an attribute of the stereotypes

«HwProcessor» and «HwComputingResource». It is

transformed to the element name.

«SwSchedulableResource» is used to annotate a class

which models a task. It is transformed to the element

task. «isPreemptible» is a MARTE attribute that is

used to specify whether a task may be interrupted. It

is transformed to the element preemptible.

The stereotypes «GaExecHost» and «Scheduler» are

used to annotate a class, which models a scheduler.

From UML/MARTE Models of Multiprocessor Real-time Embedded Systems to Early Schedulability Analysis based on SimSo Tool

205

Figure 2: ACCELEO template for MARTE to XML transformation.

Figure 3: ECLIPSE plugin for automatic MARTE to XML transformation.

These stereotypes are transformed to the element

sched. The attribute «otherSchedPolicy» is used

to specify the name of the scheduling policy. It

is transformed to the element class. To automate

any model to text transformation from MARTE to

XML model, we have transformed the implemented

ACCELEO code into an ECLIPSE plugin (Figure3).

5 CASE STUDY

To evaluate the proposed methodology for early

schedulability analysis, we have considered the Real-

Time CORBA avionics application (Madl, 2009). The

system application is composed of eleven periodic

and independent tasks such that every task is charac-

terized by a WCET (Worst Case Execution Time), a

period, a priority and a deadline (Table 2).

These parameters will be used as input for the schedu-

lability analysis step. Thus, they have been speciﬁed

in the MARTE model of CORBA system.

These properties must be entered in the tasks classes

annotated through «SwSchedulableResource» stereo-

type. Other parameters, such as the scheduling type

(static or dynamic scheduling), must also be speci-

ﬁed to handle this step. The considered application

is mapped to a preemptive execution platform com-

posed of three identical processors running at 4GHz

using 6GB three-channel RAM. The modeling of the

studied system using UML/MARTE is exposed via

Figure 4. As speciﬁed in Table 2, Gps task is peri-

odic, which is indicated through the attribute «type:

ICSOFT-EA 2016 - 11th International Conference on Software Engineering and Applications

206

Table 2: Tasks parameters for CORBA application.

Task WCET Deadline Period Priority

Gps 21 100 5 41

Airframe 53 100 15 42

Pilot_waypoints 37 100 5 43

Routes 18 100 10 44

Display_device 26 150 10 47

Af_monitor 33 120 5 49

Nav_display 14 80 5 48

Nav_steering 69 100 5 46

Navigato _navsteering _points 42 150 10 45

Pilot _control 43 100 20 50

Tactical _steering 38 80 10 51

Figure 4: The static view of the CORBA application.

ArrivalPattern» that is set to «periodic(5,ms)».

The priority of Gps task is 41, which is spec-

iﬁed through the attribute «priority:NFP_Integer

[0..*] =priority». The deadline and period of

Gps task are respectively speciﬁed through «dead-

line:NFP _Duration[0..*] = deadline» and «pe-

riod:NFP_Duration[0..*]= period» such that the dead-

line and the period are respectively set to (100,ms)

and (5,ms).

The activation date of Gps is set to 0 ms through «ac-

tivationDate:NFP_ Duration [0..*]=(0,ms)». The re-

quired scheduling type is static, which is mentioned

by setting the value of «isStaticSchedulingFeature»

to «true». The used scheduling algorithm is EDZL

(Lee and Shin, 2013). Using our ECLIPSE plugin for

MARTE to XML transformation, the MARTE model

for CORBA application has been transformed into an

XML ﬁle that will be entered to SimSo tool. Figure

5 shows the contents of the generated XML ﬁle under

ECLIPSE. The generated XML ﬁle has been entered

From UML/MARTE Models of Multiprocessor Real-time Embedded Systems to Early Schedulability Analysis based on SimSo Tool

207

Figure 5: Generated XML ﬁle after a M2T transformation.

Figure 6: Simso tool schedulability analysis report.

into SimSo for schedulability analysis of CORBA

application. Results have shown that the system is

schedulable (Figure 6).

6 CONCLUSIONS

Throughout this paper, we have proposed an MDE-

based automatic schedulability analysis at early de-

sign stages. Based on a M2T transformation from

MARTE model to SimSo tool meta-model, prop-

erties of the studied system were translated from

MARTE model to an XML ﬁle representing the input

of SimSo. A main key of this proposal is that is sup-

ports analysis of systems, which are scheduled using

semi-partitioned and global scheduling approaches.

ICSOFT-EA 2016 - 11th International Conference on Software Engineering and Applications

208

REFERENCES

(2015). Cpn tools http://cpntools.org/.

Abdeddaïm, Y., Chandarli, Y., Davis, R. I., and Masson, D.

(2014). Schedulability analysis for ﬁxed priority real-

time systems with energy-harvesting. In Proceedings

of the 22Nd International Conference on Real-Time

Networks and Systems, RTNS ’14, pages 311:311–

311:320, New York, NY, USA. ACM.

Chéramy, M., Hladik, P.-E., and Déplanche, A.-M. (2014).

Simso: A simulation tool to evaluate real-time multi-

processor scheduling algorithms. In The 5th Interna-

tional Workshop on Analysis Tools and Methodologies

for Embedded and Real-time Systems, WATERS.

Dorin, F., Yomsi, P. M., Goossens, J., and Richard, P.

(2010). Semi-partitioned hard real-time scheduling

with restricted migrations upon identical multiproces-

sor platforms. CoRR, abs/1006.2637.

ECLIPSE (2008). Acceleo http://wiki.eclipse.org/acceleo.

Gonzalez Harbour, M., Gutierrez Garcia, J., Palen-

cia Gutierrez, J., and Drake Moyano, J. (2001). Mast:

Modeling and analysis suite for real time applications.

In Real-Time Systems, 13th Euromicro Conference on,

2001., pages 125–134.

HadjKacem, Y., Mahfoudhi, A., Magdich, A., Karamti, W.,

and Abid, M. (2012). Using mde and priority time

petri nets for the schedulability analysis of embed-

ded systems modeled by uml activity diagrams. In

the 19th Annual IEEE International Conference and

Workshops on the Engineering of Computer Based

Systems (ECBS), pages 316–323.

Hagner, M. and Huhn, M. (2008). Tool support for a

scheduling analysis view. In The workshop "Model-

ing and Analysis of Real-Time and Embedded Systems

with the MARTE UML proﬁle" at DATE’08 (Design,

Automation & Test in Europe).

Henia, R., Hamann, A., Jersak, M., Racu, R., Richter, K.,

and Ernst, R. (2005). System level performance anal-

ysis - the symta/s approach. Computers and Digital

Techniques, IEE Proceedings -, 152(2):148–166.

Jensen, K. E. (2009). Schedulability analysis of embed-

ded applications modelled using MARTE. PhD thesis,

Technical University of Denmark.

Lee, J. and Shin, I. (2013). Edzl schedulability analysis in

real-time multicore scheduling. IEEE Transactions on

Software Engineering, 39(7):910–916.

Madl, G. (2009). Model-based Analysis of Event-driven

Distributed Real-time Embedded Systems. PhD the-

sis, Long Beach, CA, USA.

Magdich, A., Hadj Kacem, Y., Mahfoudhi, A., and Abid,

M. (2013a). Reducing uml/sam modeling view size

for schedulability analysis. In The 1st IEEE interna-

tional conference on Computer Applications Technol-

ogy (ICCAT). IEEE.

Magdich, A., Kacem, Y. H., and Mahfoudhi, A. (2013b).

Extending uml/marte-grm for integrating tasks migra-

tions in class diagrams. In Lee, R. Y., editor, 11th In-

ternational Conference on Software Engineering Re-

search, Management and Applications SERA (selected

papers), volume 496 of Studies in Computational In-

telligence, pages 73–84. Springer.

Magdich, A., Kacem, Y. H., Mahfoudhi, A., and Abid, M.

(2012). A MARTE extension for global scheduling

analysis of multiprocessor systems. In the 23th IEEE

International Symposium on Software Reliability En-

gineering (ISSRE), pages 371–379.

Medina, J. L. and Cuesta, A. G. (2011). From composable

design models to schedulability analysis with uml and

the uml proﬁle for marte. SIGBED Rev., 8(1):64–68.

Mraidha, C., Tucci-Piergiovanni, S., and Gerard, S. (2011).

Optimum: A marte-based methodology for schedula-

bility analysis at early design stages. SIGSOFT Softw.

Eng. Notes, 36(1):1–8.

Naija, M., Ahmed, S. B., and Bruel, J. (2015). New

schedulability analysis for real-time systems based on

MDE and petri nets model at early design stages. In

ICSOFT-EA 2015 - Proceedings of the 10th Interna-

tional Conference on Software Engineering and Ap-

plications, Colmar, Alsace, France, 20-22 July, 2015.,

pages 330–338.

OMG (2002). Uml proﬁle for schedulability, performance

and time.

OMG (2008). A uml proﬁle for marte: Modeling and anal-

ysis of real-time embedded systems.

Ramirez, A. D., Valenzuela, D. K. O., and Mejía-Alvarez, P.

(2012). A multiprocessor real-time scheduling simu-

lation tool. In 22nd International Conference on Elec-

trical Communications and Computers, CONIELE-

COMP 2012, Cholula, Puebla, Mexico, February 27-

29, 2012, pages 157–161.

Rubini, S., Fotsing, C., Singhoff, F., Tran, H. N., and Dis-

saux, P. (2013). Scheduling analysis from architec-

tural models of embedded multi-processor systems.

EWiLi Workshop.

Schmidt, D. C. (2006). Model-driven engineering. IEEE

Computer, 39.

Singhoff, F., Legrand, J., Nana, L., and Marcé, L. (2004).

Cheddar: A ﬂexible real time scheduling framework.

In Proceedings of the 2004 Annual ACM SIGAda In-

ternational Conference on Ada: The Engineering of

Correct and Reliable Software for Real-time &Amp;

Distributed Systems Using Ada and Related Technolo-

gies, SIGAda ’04, pages 1–8, New York, NY, USA.

Urunuela, R., Déplanche, A. M., and Trinquet, Y. (2010).

Storm a simulation tool for real-time multiprocessor

scheduling evaluation. In Emerging Technologies and

Factory Automation (ETFA), 2010 IEEE Conference

on, pages 1–8.

Zhang, F. and Burns, A. (2009). Schedulability analysis for

real-time systems with edf scheduling. IEEE Transac-

tions on Computers, 58(9):1250–1258.

From UML/MARTE Models of Multiprocessor Real-time Embedded Systems to Early Schedulability Analysis based on SimSo Tool

209