Loading presentation...

Present Remotely

Send the link below via email or IM

Copy

Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.

DeleteCancel

Minimizing Nasty Surprises with Better Informed Decision-Ma

No description
by

Sara Hassan

on 15 June 2016

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of Minimizing Nasty Surprises with Better Informed Decision-Ma

Thank you!
Questions?

Goal
"Claim"?
Run-time testing is needed to verify the stability, resilience and NFR satisfiability
Stakeholder identification
Reliance on stakeholders
Time-sensitivity
Scalability of the method
Minimizing Nasty Surprises with Better Informed Decision-Making in Self-Adaptive Systems
Self-Adaptive?
Nasty Surprise?
M = Satisfying NFRs
D = Evidence
Unit = wow
Bigger magnitude, higher need to trigger an adaptation
"Nasty" due to cost
Informed Decision-Making?
Systematic method to:
Explore potential environment impact on adaptation rate:
Using simulated surprises
Probabilities estimated from historical data
What-if analysis to choose SAS design which:
Mitigate risks on NFRs
Balance conflicting NFRs
Pareto Analysis
Calculation of Surprise Value Distribution and
Tolerance Threshold

Multi-attribute Risk Assessment and Benefit Assessment

NFR 1?
Yes
No
NFR 2?
NFR 2?
Yes
No
Yes
No
GridStix
Wireless sensor network for flood prediction
Submits readings to a gateway
Based on plug-ins to create a component framework
NFRs
Fault tolerance: routes for data to reach the gateway
Energy efficiency: sensor power consumption
Monitorable
River depth (assumed to affect both NFRs)
Risks
Fault tolerance:
Exceeding surprise threshold
Protocol failure
Wrong plug-in design
Energy efficiency:
Exceeding surprise threshold
Long delay in passing data
Failure to report power shortages
Solution Space
For planning the path of the data:
First hop algorithm
Shortest path algorithm
For communication between sensors:
BlueTooth
WiFi
Designing new plug-ins
Calculation of Surprise Value Distribution and Tolerance Threshold
Large range for monitorable
The points indicate the need for adaptation
Shape of the graph indicates the significance of the evidence
Below the tolerance threshold goal weakened
Multi-attribute Risk Assessment and Benefit Assessment
Impact calculated by threat index.
Threat index use a vector of attribute values
Attributes here:
Node failure
Power consumption overhead
Mitigation estimated as a percentage change
Pareto Analysis
Claim:
WiFi + SP
Contribution
A
comprehensive systematic design-time
method for what-if analysis to:
To
minimize
the "nasty" surprises
Risk-aware

Conflict-aware

Goal weakening
possible
Scalable
to as many conflicting NFRs as needed (flexible representation)
Sara Hassan*
,
Nelly Bencomo^, and Rami Bahsoon*

Sara Hassan: ssh195@bham.ac.uk
, Nelly Bencomo: nelly@acm.org, Rami Bahsoon: r.bahsoon@cs.bham.ac.uk
*School of Computer Science, University of Birmingham, UK
^Aston University, Birmingham, UK
Full transcript