Transcript UrbanFlood
UrbanFlood
The Common Information Space:
A Framework for Early Warning
Systems
Bartosz Baliś, Tomasz Bartyński, Marian Bubak,
Grzegorz Dyk, Tomasz Gubała, Marek Kasztelnik
Motivation: Early Warning Systems
against natural disasters
Complex distributed systems
Mission and time-critical
Involve resource-intensive computations and data
analysis
Require compositions of resources into complex
scenarios
◦ Sensors monitoring objects and natural phenomena
◦ Apps for data analysis & simulation
◦ ICT infrastructure (hardware + software)
Example: Urban area flood
simulation
Goals
Develop a framework facilitating
development, deployment and execution
of EWSs: the Common Information Space
EWS development
◦ EWS reference model
◦ EWS development framework
EWS deployment
◦ EWS blueprints
◦ EWS-factory-as-a-service
EWS execution
◦ CIS runtime services for resource allocation, selfmonitoring, self-healing, mission-critical operation, and
urgent computing
CIS Architecture and EWS
reference model
Cracow Grid Workshop 2012
UrbanFlood
5
CIS as a system factory
System instance
System factory
invoke
System
blueprint
configurations
Service factory
Service factory
Service
Service
configurations
customize
Virtual appliance store
Cracow Grid Workshop 2012
UrbanFlood
6
Case study: Flood Early Warning
System
Protection of urban areas against floods due to
dike failures
Real-time monitoring of embankments using
wireless sensors
A casade of models for:
◦ Anomaly detection
◦ Risk assessment
◦ Impact prediction
Cracow Grid Workshop 2012
UrbanFlood
7
Dike anomaly detection
Anomaly detection using neural cloud
classification
Dike breach risk computation
Distribution of probability of dike
breach is calculated
Depending on the current water level
Flood simulation
Life Safety Model
Loss of life estimate
due to flood
Initial state of the world
Post-event summary
(red color = deceased)
Virtual Dike
Computer model of a dike
Simulation of dike behaviour
Useful for „what-if” experiments
Also for testing the real EWS
(„emergency drill”)
Flood Early Warning System
workflow
Dike
Data
flow
Control
flow
Sensor
Cabinet
Alternate path
Virtual
Dike
AnySense
Service
AI Anomaly
Detection
action
Reliability
Analysis
Abnormal?
High risk?
Life Safety
Simulation
Flood
Simulation
High
impact?
Flood EWS: implementation
Cracow Grid Workshop 2012
UrbanFlood
14
Example: Flood Simulation Part
Cracow Grid Workshop 2012
UrbanFlood
15
Conclusion
CIS contributes conceptually and technologically
in the following areas:
CIS as a factory for Early Warning Systems
◦ Proven by the implementation of the Flood EWS
CIS as a geo-ICT technology for spatial
data processing services
◦ CIS adopts guidelines of the INSPIRE directive and
leverages OGC standards for service interfaces
CIS as a runtime infrastructure for
resource-intensive mission-critical systems
Cracow Grid Workshop 2012
UrbanFlood
16
urbanflood.cyfronet.pl
dice.cyfronet.pl/cis
Cracow Grid Workshop 2012
UrbanFlood
17