Embedded Systems at Thales: the Artemis challenges from an industrial perspective

38 Pages·2010·6.53 MB·English

Embedded Systems at Thales: the Artemis
9"
0
0
2
09/ challenges from an industrial perspective
9/
0
ol
o
c
S
er
m
m
u
S
st
Arti
« Gilbert Edelin
Research Group Director
Research & Technology
Company confidential
Thales profile
Three core Defence A Worldwide Group
businesses • 68,000
employees
50 %
worldwide
Air
  • Presence in 50
Land
 
countries
Naval
 
Joint
 
Aerospace & Space Security
25 %
25 %
World leader for mission-critical information systems
2
Innovation and technological excellence
 R&D at Thales totals €2.4bn (19% of revenues)
 25,000 researchers on cutting-edge technologies
 300 inventions per year
 Over 15,000 patents
 Over 30 cooperation agreements with universities and public
research laboratories in Europe, the United States and Asia

9
0
0
2

/
9
0
/
9
0

l
o
o
h
c
S

r
e
m
m
u
S

T
S
I Copyright © Nobel Web AB 2007
T
R Photo: Hans Mehlin
A
3
Group Restricted
Thales and Embedded Systems
Thales leader in Mission Critical Systems. Most of them are Embedded,
trend toward more autonomy
Continuing need of High Performance Embedded Systems: innovative
sensor algorithms, parallel architectures, engineering from algorithm to
parallel implementation,
Common issues: life cycles, ITAR compliance

9
0
0
2

/
9
0
/
9
0

l
o
o
h
c
S

r
e
m
m
u
S

T
S
I
T
R
A
4
Group Restricted
Embedded Systems life expectancy far exceeds that of their
components
2 to 4 years
Processor
generation
10 years
Car

9
0
0
2

/
9
0
/
9
0

l
o
25 years
o
h
c
Aircraft
S

r
e
m
m
Life expectancy
u
S

T
S The reconciliation between those industrial discrepancies
I
T
R must stem from elsewhere: architecture, methods
A
5
Group Restricted
Introduction: Thales and Artemis
Artemis SRA defines 3 research areas with cross-cutting
objectives between application domains:
 Reference Designs and Architectures: high performance and
dependability in the multi-manycore era
9  Seamless connectivity, Middleware: focus on heterogeneous
0
0
2 solutions to deal with RT

/
9
0
/  Design methods & tools: beyond MDA, Domain specific solutions
9
0

l
o
o
h
c
S
The diversity of the Thales embedded applications urges to

r
e
m
increase in-house cross-cutting solutions between domains
m
u
S

T
S
I
T
R
A
6
Group Restricted
Avionics Systems have gone to “generic”processors
A 380 : 80 Mbytes

9
0
0 Hard coded logics
2 Concorde 1970

/
9
0
/
9
0

l
o
o
h
c
S

r
e
m
m
u
S

T
S
I
T Software has become “intensive” to bring more functional value
R
A
7
Group Restricted
New hard real-time needs keep on appearing
 growing automation (transportation, industrial processes, ...)
 E.g. auto-Pilots & Flight Control, Fighters Terrain Following,
Unmanned Air Vehicles …
9  new real-time media/interaction demanding QoS1 (video, Positioning..)
0
0
2
 E.g. Cameras & SAR radar video for observation,
/
9
0
In-Flight Entertainment Video on Demand,
/
9
0
anti-collision Systems (TCAS)…
l
o
o
h
c
S

r
e
m
m
u
S

T
S 1 QoS: Quality of Service
I
T
R
A
Source : Jean-Luc Voirin
8
Aerospace Division Group Restricted
Some (new) challenges for future aerospace systems

9
0
0
2

/
9
0
/
9
0

l
o
o
h
c
S

r
e
m
m
u
S

T
S
I
T
R
A
Source : Jean-Luc Voirin
9
9
Aerospace Division Group Restricted
All ES are impacted by architectural breakthroughs
Technology limitations: perf. by
Processor parallelism no more by frequency =>
Multi-processing
Power disruption in programming model, long
term research challenges
Processor
PowerQUICC II specialization
Architectural
break point
Domain oriented architectures: eg with
predictable performance to control the
timeliness in RT critical applis,
dynamic reconfiguration for adaptive,
RISC era
distributed critical architectures
(multilevel RT composability)
CISC era
Intel 8086
Time
1980 1990 2000 2015-2020 (?)
A major architecture disruption: multiprocessing and specialization
will have a strong impact on software
10