institute of computer and communication network engineering multi-dimensional robustness...
Post on 21-Dec-2015
219 views
TRANSCRIPT
Institute of Computer and Communication Network Engineering
Multi-dimensional Robustness Optimization of Embedded Systems
& Online Performance Verification
Arne Hamann
Steffen Stein
Rolf Ernst
Institute of Computer and Communication Network Engineering
Part I:Multi-dimensional Robustness Optimization of Embedded Systems
Arne Hamann
Rolf Ernst
3
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Outline
• System property variations
• Sensitivity Analysis
• Stochastic Multi-dimensional Sensitivity Analysis
• Robustness Metrics– Hypervolume calculation– Minimum Guaranteed Robustness (MGR)– Maximum Possible Robustness (MPR)
• Experiments
4
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
System Property Variations
• Why do system property variations occur?– Specification changes, late feature requests,
product variants, software updates, bug-fixes
• Robustness to property variations– decreases design risk, and increases system
maintainability and extensibility
• Property variations can have severe unintuitive effects on system performance
• Sensitivity analysis: achieve robustness without on-line parameter adaptation
5
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Problem Formulation
• Find fixed parameter configuration that …• … maximizes system robustness w.r.t.
changes of several properties• Robustness = the system can sustain
property variations without severe performance degradation
• Not included: dynamic parameter adaptations (ongoing work submitted to EMSOFT 2007)
6
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Stochastic Sensitivity Analysis (1)
• Problem of exact sensitivity analysis approaches: computational effort grows exponentially with number of considered dimensions
• Solution: scalable stochastic analysis able to quickly bound system sensitivity
7
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Stochastic Sensitivity Analysis (2)
• Sensitivity analysis formulated as multi-objective optimization problem
Pareto-front of optimization task corresponds to sought-after sensitivity front
• Use multi-criteria evolutionary algorithms to approximate sensitivity front– E.g. SPEA2 (ETH Zurich): diversified sensitivity
front approximation through Pareto-dominance based selection and density approximation
8
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Creation of the Initial Population
• Creates a certain number of points representing a first approximation of sensitivity front
• Uses 1-dim sensitivity analysis– to bound the search space in each dimension
(bounding hypercube)– to generate points representing the extrema of the
sought-after sensitivity front
• Randomly place the rest of the initial points in bounding hypercube
9
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
6
6,5
7
7,5
8
8,5
9
9,5
10
10,5
11
9 11 13 15 17 19 21 23 25 27
6,5
10,85
26,210Property 1
Pro
pert
y 2
Bounding Box
Initial Population - Example
10
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Bounding the Search Space (1)
• Idea: bound search space containing the sought-after sensitivity front– Bounding working Pareto-front F n
• evaluated Pareto-optimal working points
– Bounding non-working Pareto-front F nw
• evaluated Pareto-optimal non-working points
• Bounding Pareto-fronts can be used to derive multi-dim. robustness metrics (later)
11
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Bounding the Search Space (2)
• Space between bounding Pareto-fronts is called relevant region
• Variation operators use algorithm ensuring that generated offsprings (points) are situated in the relevant region– Below bounding non-working Pareto-front– Above bounding working Pareto-front
Efficiently focuses exploration effort
12
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Bounding the Search Space (3)
6
6,5
7
7,5
8
8,5
9
9,5
10
10,5
11
9 11 13 15 17 19 21 23 25 27
6,5
10,85
26,210Property 1
Pro
pert
y 2
Bounding Box
13
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Front Convergence Mutate (1)• Heuristic operator adapted to optimization
problem• Strategy:
– Determine X closest points on opposite Pareto-front
– Choose randomly one of these points– Place offspring point randomly on straight line
connecting the parent point and the chosen random point
• Increases convergence speed of the bounding Pareto-fronts
14
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Front Convergence Mutate (2)
6
6,5
7
7,5
8
8,5
9
9,5
10
10,5
11
9 11 13 15 17 19 21 23 25 27
6,5
10,85
26,210Property 1
Pro
pert
y 2
Bounding Box
15
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Front Convergence Mutate (3)
6
6,5
7
7,5
8
8,5
9
9,5
10
10,5
11
9 11 13 15 17 19 21 23 25 27
6,5
10,85
26,210Property 1
Pro
pert
y 2
Bounding Box
16
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Hypervolume Calculation• Hypervolume as basis of the proposed
robustness metrics• Hypervolume is defined in a given
hypercube and associated to a point set• Two different notions of hypervolume
– inner hypervolume : Volume of space Pareto-dominated by the given points inside the given hypercube
– outer hypervolume : Volume of space Pareto-dominated by all points not Pareto-dominating any of the given points
17
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Hypervolume Calculation (2)• 2D-case
– inner hypervolume: lower step function– outer hypervolume: upper step function
4
6
8
10
12
14
16
18
20
14 16 18 20 22 24 26 28 30
Bounding Box[15,28]x[6,18]
(15,18)
(18,16)
(20,12)
(26,10)
(28,6)
( )= 66λ-( )= 100λ+
18
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Robustness Metrics
• Given a set of properties …
• … use stochastic sensitivity analysis to derive upper and lower robustness bounds– Minimum Guaranteed Robustness (MGR)
• Defined as inner hypervolume of the bounding working Pareto-front F w
– Maximum Possible Robustness (MPR)• Defined as outer hypervolume of the bounding
non-working Pareto-front F nw
19
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Robustness Metrics (2)
Obviously: MGR <= Real Robustness <= MPR
6
6,5
7
7,5
8
8,5
9
9,5
10
10,5
11
9 11 13 15 17 19 21 23 25 27
6,5
10,85
26,210Property 1
Pro
pert
y 2
Bounding Box
MGRMPR
20
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Robustness Exploration
• Idea: Pareto-optimize MGR and MPR
• Advantages– Stochastic sensitivity analysis is scalable
Little computational effort necessary to reasonably bound robustness potential of given configuration
– In-depth analysis can be performed once interesting configurations are identified (i.e. high MGR or high MPR)
Perfectly suited for robustness optimization
21
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Example System
• Distributed embedded system
• 4 computational resources …
• …connected via CAN bus
• 3 constrained applications– SensAct
– SinSout
– CamVout
22
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Approximation Quality (1)
• Approximation after 100 evaluations (20 sec)
• MGR = 2447• MPR = 2937
• Approximation after 200 evaluations (40 sec)
• MGR = 2580• MPR = 2813
23
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Approximation Quality (2)
• Approximation after 300 evaluations (60 sec)
• MGR = 2632• MPR = 2777
• Result using exact sensitivity analysis (85 sec)
• MGR = 2585• MPR = 2826
24
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
3D - Robustness Maximization
Original configuration
Optimized configuration
25
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Integration of New Functionality
• Integration of a fourth application with lowest priorities
• What combinations WCET T9 and WCCT C6 are feasible?
• Is there optimization potential?
• Idea: initially assume WCET T9 and WCCT C6 equal zero
T9
C6
Sens2
Sink
26
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Integration of New Functionality (2)
0
5
10
15
20
25
30
0 5 10 15 20 25 30 35 40
Original
Optimized
WC
CT C
6
WCET T9
• Areas below the curves represent feasible systems
Institute of Computer and Communication Network Engineering
Part II:Online Performance Verification
Steffen Stein
Rolf Ernst
28
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Outline
• Motivation
• Framework Architecture
• In Detail: Global Analysis Layer
• System Setup
• Approach to Analysis Control
• Experimental Results
29
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Future Challenges
• In-Field updates• Run-time Reconfigurations• 90% of Innovation in Software • Networked Systems
Not Manageable
at Design-Time!
0
5
10
15
20
25
30
0 5 10 15 20 25 30 35 40
Original
Optimized
Engine Control SW
Driver Assistance
Multimedia Service
30
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Approach
• Generally Speaking:– Make Systems clever enough to handle Integration
Problem themselves
• Here: Timing Properties• ToDo
– Gather performance Data during runtime– Evaluate/ Optimise online– Feed Results back into running Systems
• Result: Evolving Systems
31
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Architecture: Organic Computing
• Single Instance• Multiple Instances• Multiple collaborating
Instances• Layered approach
System under Observationand Control (SuOC)
observer controller
observes controls
reports
selects observation model
Goals/Design Rules
Source: Towards a generic observer/controller architecture for Organic Computing, U. Richter, M. Mnif, J. Branke, C. Müller-Schloer, H. Schmeck, INFORMATIK 2006 -- Informatik für Menschen
32
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Analysis EngineAnalysis EngineObserver ControllerObserver Controller
Heterogeneous Networked Embedded System (SuOC)
Analysis Engine
Control Plane
Local Layer
Use resourcesGather data Adjust settings
Observer ControllerData
Exchange
Global Analysis Layer
Global Controller Layer
Global Observer Layer
Self-Organisation
Control Framework
Self-Organisation
33
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Distributed Setup
ARMDSP
uCPPC
CAN
Real System
T2
T5
T1
T6
S2
S1
S4
S3
T8
T0
T7
T9
T3
T4
Global Model
34
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Analysis Control
T1
T3
T2
T4
T1
T6
Net
wor
k T
unne
l
Distributed Analysis Control
T1
T3
T2
T4
T1
T6
Analysis Control
Dofor all not up-to-date Resources
Analyseend
Until all Resources are up to date
While (true)if Resource invalidated
analyse Resourceend
end
35
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Performance of trivial Approach
0
100
200
300
400
500
600
700
800
900
1000
0 20 40 60 80 100
Live
offline
System size (# tasks)
# a
naly
sis
ru
ns (
resou
rce level)
36
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Problem
T1
T4
T2
T5
S1
S2
T3
T6
T7 T8S3 T9
Exponential increase in number of necesary Analysis runs
Solution: Caching
37
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Performance with caching
0,00
20,00
40,00
60,00
80,00
100,00
120,00
140,00
160,00
180,00
0,00 20,00 40,00 60,00 80,00 100,00 120,00
Live
Offline
System size (# tasks)
# a
naly
sis
ru
ns (
resou
rce level)
38
Arne Hamann, Steffen Stein, IDA, TU Braunschweig
Conclusion
• Distributed Performance Analysis implemented
• Suitable as evaluator for online performance control / optimization
• Future Work: From System observations to analysable Model.