The document discusses challenges in analyzing heterogeneous component-based software systems, focusing on safety monitoring and control systems. It emphasizes the need for a model-driven information flow analysis approach to ensure that sensor signals reach the correct actuators for certification purposes and proposes building a homogeneous model of the system. Limitations are noted, including issues with bit manipulation and the need for improvements in abstraction and visualization for comprehensibility.

1. Crossing the Boundaries while Analyzing
Heterogeneous Component-Based
Software Systems
Amir Reza Yazdanshenas
Leon Moonen
2011/09/28
ICSM 2011

2. Safety Monitoring and Control System
2

3. Safety Monitoring and Control System
Sensor Actuator
Sensor logic Actuator
Input Output
Sensor Actuator
3

4. Safety Monitoring and Control System
logic
Input Output
4

5. Case Description
Start with clear More costumers: More costumers:
design & documents new requirements scalability
Input1 Output1
Input2
A A
Output2
• Inhibit Input1
A A Output1
D D
Input3
D D
Output3 • Override
Input4 Output4
Input5
D D
Output5
• Suppress
D D
Input6
D D
Output6 • Acknowledge CascIn CascOut
Input7 Output7
A D
D Output8
Output9 A A
D
Similar to digital circuit CascIn
CascOut
design components!
Workaround:
• Cascading modules
• Voting modules
5

6. Case Description…
More costumers:
reusing decisions
6

7. Case Description…
 Potentially unlimited number of configurations
– No default/standard configuration of components!
– Separate configuration for each installation
– (examples omitted due to NDA)
7

8. Problem Statement
logic
Comp. 1 Comp. 3
Input Output
Comp. 2 Comp. n
8

9. Problem Statement
Do sensor signals reach the correct output actuators?
logic
Input Output
For certification purposes,
can we provide source based evidence?
9

10. Tracking Information Flow
“find source based evidence that
signals from sensors trigger the correct
actuators”
is there information flow from
the desired sensors to the selected actuator?
are the desired sensors (input ports)
part of the backward program slice
for the selected actuator (output port)?
10

11. Heterogeneous Systems
 Deployed system is not just set of components
– actual behavior depends on composition &
configuration
– literature focuses on analysis of homogeneous
systems
⚡ Existing slicing tools are language specific
⚡ no support for “external” artifacts
11

12. Challenge #1:
void main() { void main() { void main() { void main() {
int sum, I;
while ( i<11 ){
sum = add(sum)
int sum, I;
while ( i<11 ){
sum = add(sum)
int sum, I;
while ( i<11 ){
sum = add(sum)
✗ int sum, I;
while ( i<11 ){
sum = add(sum)
i = add(i, 1); i = add(i, 1); i = add(i, 1); i = add(i, 1);
… … … …
✗
✓
✗
sensor.c input.c voter.c output.c
© 2011 Leon Moonen Model-driven Information Flow Analysis to Support Software Certification - NECSIS Seminar at Queen's (2011/06/27) 12

13. Shared Memory Communication
main
cause & effect
cause matrix effect
input A 1 2 3 j
output B
2
3
effect
output C
input B cause i
output D
effect
13

14. Challenge #2:
output#1
sensor#1
input#1
output#2
sensor#2
input#2 output#3
sensor#N
output#4
14

15. Challenge #2:
output#1
sensor#1
input#1
output#2
sensor#2
input#2 output#3
sensor#N
output#4
15

16. Our solution: Build a Homogeneous
Model of the System
16

17. KDM: Flexible and Extensible
17

18. SDG: ICDG + CDGs
18

19. Conclusion
void main() { void main() { void main() { void main() {
int sum, I; int sum, I; int sum, I; int sum, I;
while ( i<11 ){ while ( i<11 ){ while ( i<11 ){ while ( i<11 ){
sum = add(sum) sum = add(sum) sum = add(sum) sum = add(sum)
i = add(i, 1); i = add(i, 1); i = add(i, 1); i = add(i, 1);
… … … …
sensor.C input.C voter.C output.C
19

20. Conclusion
output#1
sensor#1
voter#1
input#1
output#2
sensor#2
voter#2
input#2 output#3
sensor#N voter#3
output#4
20

21. Conclusion
✓ ✓
✓ ✓
✓
✓ output#1
✓
✓ sensor#1 ✓ ✓
✓
voter#1 ✓
input#1 ✓
output#2
✓
sensor#2
✓
voter#2
✓ ✓
input#2 ✓ ✓ output#3
✓
✓
✓ sensor#N ✓ voter#3
output#4
21

22. Precision & Scalability Tests
 Identical results with CodeSurfer on an
example program
 Kongsberg code base:
22

23. Precision & Scalability Tests
 Identical results with CodeSurfer on an
example program
 Kongsberg code base:
23

24. Limitations
 Bit manipulation: no precise information flow
– granularity limitation in Codesurfer implementation
 Implemented for C with proprietary composition
– experiment with other languages, e.g. Java and other
composition languages/frameworks
Future Work
 Abstraction and visualization
– improve comprehensibility of results
– present the result in multiple abstraction layers
– separate Intra- and Inter- component information
flows
24

25. Thank you!
Comments/Questions?
25