16
MultiPARTES Towards Model-Driven Engineering for Mixed-Criticality Systems: MultiPARTES Approach A. Alonso , C. Jouvray, S. Trujillo, M.A. de Miguel, C. Grepet, J. Simó WICERT 2013, March 22nd

MultiPARTES

  • Upload
    krista

  • View
    24

  • Download
    0

Embed Size (px)

DESCRIPTION

MultiPARTES. Towards Model-Driven Engineering for Mixed-Criticality Systems: MultiPARTES Approach A. Alonso , C. Jouvray, S. Trujillo, M.A. de Miguel, C. Grepet, J. Simó WICERT 2013, March 22nd. Motivation and goals. - PowerPoint PPT Presentation

Citation preview

Page 1: MultiPARTES

MultiPARTES

Towards Model-Driven Engineering for Mixed-Criticality Systems: MultiPARTES Approach

A. Alonso, C. Jouvray, S. Trujillo, M.A. de Miguel, C. Grepet, J. SimóWICERT 2013, March 22nd

Page 2: MultiPARTES

2

Motivation and goals

Modern electronic systems used in industry (avionics, automotive, etc) combine applications:

– with real-time (hard and soft) & no real-time requirements– with different levels of security– that can be independently qualified (certified)

This trend can imply increased validation and certification costs

This extra cost can be reduced by:– Isolate the execution of different applications– Methodology with higher abstraction level

Page 3: MultiPARTES

3

MultiPARTES goals

MultiPARTES– FP7, in the area of ICT– http://www.multipartes.eu

MultiPARTES goals:– To develop a multicore platform virtualization layer for

critical and secure embedded systems.– To propose a methodology to enforce the rapid

development of new applications based on partitioned systems

– To develop methods and tools to support the application development

Page 4: MultiPARTES

4

Approach Overview

Design an embedded system composed by a set of applications

Execute this embedded system in a:– Partitioned execution environment– Multi-core platform

Supported by a virtualization layer

Heterogenous system– Different processors (+/- predictable, +/- powerful)

VLVL VLVL

ProcessorProcessor ProcessorProcessor

Page 5: MultiPARTES

5

Hypervisor

Hypervisor based system permits to build partitioned systems where partition:

– Are temporal & spatial isolated

– Use the appropriated OS for each application

– Execute mono-core OSs in a multicore platform

Page 6: MultiPARTES

6

XtratuM Hypervisor

Open source bare-metal hypervisor for critical real-time partitioned systems

Uses para-virtualization techniques– Strong temporal isolation: fixed cyclic scheduler– Strong spatial isolation: every partition is executed in

processor user mode and does not share memory. – Robust communication mechanisms (ARINC ports)– Robust error management via the Health-Monitor– Devices can be directly managed by partitions. Shared

devices can be organized in a IOServer– Resources are allocated statically through a Configuration

file (XML)

Page 7: MultiPARTES

8

www.multipartes.eu

Development process

Methodology to enforce the rapid development and production of new applications based on partitioned systems

High level system model:Set of applications and interconnections •Mixed-criticallity•High level model of hardware resources:Types of devices and propertiesResources management models

Application constraints:criticality, time, device usageresource availability

Page 8: MultiPARTES

9

MDE Approach

Model Driven Engineering (MDE) approach– facilitates to bridge the gap between design issues and

partitioning concerns– Models are the main development artifacts– Annotation for non-functional properties

Approach based on several metamodels

Page 9: MultiPARTES

10

Toolset

Multipartestoolchain withUML profiles

Page 10: MultiPARTES

11

Platform model

This model defines the relevant properties of the platform

Page 11: MultiPARTES

12

HW Platform Metamodel

Page 12: MultiPARTES

13

Application model

Described using UML Enriched with non-functional annotations

– Criticality requirements: in terms of safety levels– Time requirements: based on UML MARTE– Resource needs requirements: for guaranteeing QoS– Device usage requirements

Page 13: MultiPARTES

3

Partitioning

Information to be used for partitioning Components that must be in the same partition (App) Level of criticality (App) Time requirements granularity (App) Requirements on OS (App) Components that must be executed on a processor (Par) Hardware platform (HW)

Defining a metamodel for the partitioning model

Requirements on partitioning that relates the platform and the application.

Deployment model for code and configuration generation

3

Page 14: MultiPARTES

15

Reference platform

A heterogeneous platform based on– One Atom multicore processor (general platform)

• Less critical applications• Higher computation capabilities• General Purpose OSs

– Two LEON3 synthesized in FPGA • Higher predictability• More critical applications• Lower computation capabilities• Real-time OSs

Page 15: MultiPARTES

16

Conclusions

Mixed-criticality systems are required for industry and are a challenging topic

MultiPARTES approach– Based on an XtratuM: hypervisor that provides spatial and

time isolation– MDE: basis to facilitate system development– Tool framework:

• Allows to define non-functional requirements related with partitioning

• Time and safety requirements are validated• Generation of code and configuration files

Page 16: MultiPARTES

17

Questions?

More information:http://www.multipartes.eu