Transcript Design of a Reliable Resilient Real

CONFIDENTIAL
Reliable Resilient Real-time
Flood Warning Network
James Logan, OneRain,
CONFIDENTIAL
Flood Warning
Summary
• What is in a real-time monitoring
network?
• How can it be made it reliable and
resilient?
– Identify single points of failure, options for
mitigating single points of failure
– Based on risk and cost, implement
redundancy where it makes sense
What make a real-time
monitoring network?
What does it mean to be reliable and
resilient?
• Reliability is the achieved outcome
– It is up every time it is needed
– Data is available anytime/anywhere
• Resiliency lies in the design and
maintenance
• Design minimizes single points of failure
• Strong maintenance practices and tracking help
achieve accountable performance
Telemetry
• Affordable and sustainable
– Power management, cost, vandalism issues
• Timely delivery
– Flash flooding requires real-time data
• Reliable paths
– RF and network
• Redundant paths increase data reliability
Analyze for points of failure
• Gauge Site Failures
– Is redundancy required here?
– If you lose a sensor, will your operations
fail?
• Repeater Failures
– If you lose a repeater, what data are lost?
Analyze for points of failure
• Base Station Failures
– Power, antenna, receiver, decoder; Will one
failure take the system down?
– Connectivity for dissemination; How many
users are impacted? Will alarms go out?
Reliable networks are designed to
avoid single points of failure
• Move single points of failure out to gauges
• Redundant telemetry at repeater sites
– Independent receive/transmit, different data
channels (ALERT, satellite, IP, cell, ALERT2™)
• Redundant data receive sites
– Geographically distributed, diverse mechanisms
Mitigating missing rain gauges
• Importance of rain gauges
– If a gauge does not participate, you have lost that
area of rainfall monitoring, could be a critical
catchment
• GARR lessens impact of lost rain gauges
– Gauges used to calibrate radars superior spatial
coverage
– No loss of information at gauge that failed, less
certainty of overall accuracy than with the gauge
Rain gauges vs. gauges + radar
Gauge-only rainfall
estimates
Gauge-adjusted radar rainfall
estimates
Reliable networks are well-maintained,
their performance held accountable
• Solid, standardized maintenance practices
– Routine/proactive scheduled maintenance
– Good trouble recognition and troubleshooting
– Daily analysis of data, statistical tools
• Performance accountability
– Daily/monthly/annual reporting to show
sensor, network availability
Data collection & dissemination
• Reliable/redundant receive points
– Safe archiving once data arrive
– Lost gauge data are irreplaceable
• Accessibility
– Anytime/anywhere access
– Usability
• Appropriate security
– Authorized users? What should the public see?
Real life examples
Los Angeles County
• Redundant repeater path
• ALERT RF and StormLink™ Satellite
Concentrators at Repeaters
• Redundant base stations
• DIADvisor™ and Contrail® Web for Base
Station Redundancy
Denver Urban Drainage and
Flood Control District
• Redundant repeater path
• ALERT RF and ALERT2™ at Repeaters
• Redundant base stations
• Mission critical customer agencies of UDFCD
have their own base station receiver and
software as backup to UDFCDs website
Monterey County
• Repeater Redundancy
• ALERT RF and StormLink™ IP Concentrator
at repeater
• Base Station Redundancy
• Local DIADvior™ Base Station and Contrail®
Web for base station redundancy
Napa County
• Repeater Redundancy
• ALERT RF and StormLink™ Satellite
concentrator at repeater
• Base Station Redundancy
• Local DIADvisor™ Base Station and
Contrail® Web for base station redundancy
Summary
• Identify single points of failure
• Investigate options for reducing single
points of failure
– Network backbone
– Base Stations
• Based on risk and cost, implement
redundancy where it makes sense
Thanks!
Questions?
CONFIDENTIAL
StormLink™ Satellite
Telemetry
•
•
•
•
L – Band
20 second latency
Independent channel
Independent
infrastrucure
– Contrail® Web
MSAT Satellite
X.25
StormLink™ rainfall example
Downlink
X.25
Reston, VA
Radar Data gauge Data
TCP/IP
Customized Products
VPN Tunnel
Internet
XML
http
Remote Site
Client Application (DIADvisor™, other)
Contrail® client
Network Evaluations
•
•
•
•
•
Harris County, TX
Louisville MSD, KY
Overland Park, KS
Clark County, NV
Denver Urban Drainage
and Flood Control District
• Southern CA ALERT
Network (SCAN)
• Maricopa County, AZ
• Entergy Corporation
Monitoring components
• Site design issues, for example:
– Good capture by rain gauge?
– PT in the water at low levels?
– Vulnerability to high flows?
• Representative monitoring, for example:
– Rainfall – are gauge data used alone or to
calibrate radar (don’t need as many gauges
with GARR)?
BOR/BIA Dam Safety Project
• Since 2004
• 130+ high risk, high
hazard dams
• Contrail® Web with
automated alarms &
notifications
• Supporting National
Monitoring Center,
staffed 24/7
Louisville/Jefferson County MSD
• Since 2003
• Real-time control support
• MSD, USGS and
METAR gauges
• OneRainware™ GARR
– Real-time, 4-hour forecast
• Contrail® Web with
automated alarms &
notifications
X.25
MSAT Satellite
StormLink™ rainfall example
Downlink
X.25
Reston, VA
TCP/IP
VPN Tunnel
Internet
XML
http
Remote Site
Client Application (DIADvisor™, other)
Contrail® client