IOT Use Cases Aug 2014 Intel Campus
Oct 21, 2014
IOT Use Cases Aug 2014
Intel Campus
Social Signal Mechanical signal
What e ll o er i e t 30 mins
1. 12 IOT Technical Design Patterns + Biz Use cases
2. 5 Key Learnings
3. 7 Emergent Transitions which play itself out
Landing gear becomes an addressable NODE !
Machines meet Software !
• The Ulti ate U io … Software + Ph si al i ter o e … walls are falling
• Value moving from Machines S/W
• Machines as a Service ( MAAS )
• Intelligence can live anywhere
• Edge
• Command centre
• Anywhere between machines controller and universal network level
Scenario-1 : Event sink
Massive sensor event ingestion
Flow sensors
Steam pressure
sensors
Temperature
sensors
Status sensors
Scenario-2 : Real time event Monitoring
States + Alarms
Scenario-3 : Vertical spe ifi Chai of thought a al sis
Chain of thought analysis
Scenario-4 : Ooops is my Asset down ?
Detecting Breakdowns !
Scenario-5 : Is it reall ad ? Tria gulati g false alarms
• A smoke alarm goes of on second floor
• It could be a smoker doing it in isolation
• GCC wants to verify it
• They look at the status of adjacent sensors on same floor on different floor
TRIANGULATE !
Scenario-6 : I a t to Dig deep i to a asset
Digging Deeper into Strategic asset sensors
Discovering buried cause and effects
Scenario-7 : I ha e a hu h – Hypothesis testing of domino effect
Scenario-8 : Whi h are the alar hotspots
Geospatial dimension
Scenario-9 : Adverse event Message forensics
E a i e u tou hed messages !
Scenario-10 : Looking for a needle in a haystack
Thin Slice !
X
1
2
3
4
Day-1 9:12 am
Fuse blowout
Day-6 3:15 pm
Line
Maintenance
Inspection
Day-9 10:15 pm
Transformer or
Eco system device
recalibration
Day-11 3:15 pm
Three 2 sigma Voltage
Surge Event occurred
Within 24 hours
Day-12 3:15 pm
Power Outage
Breaks down
X
X
X
X
X
X
X
X
X
X
X
Scenario-11 : Diagnosing Asset outcomes
Scenario-12 : Rank order Predictors of process or asset outcomes
So what are the key learnings ?
1. Bridging Engineering mental model Software mental model
ENGINEER DATA GUY
0
2000
4000
6000
8000
10000
12000
14000
16000
1 1 1 1 1 2 10 100 120 130 140 1000
Wri
te S
pe
ed
Number of threads
Write speed with the number of threads
Cassandra HBase
2. Stress test workloads
3.Beware of data blind spots for modelling Asset Outcomes
What are the 7 irreversible transitions have been set in motion ?
Transition-1 :
From Atoms to Bytes
Mech intensive Intelligence intensive
Transition-2 :
From Uncoupled to Tightly coupled IT-OT
Transition-3 :
From Fragmentation to Standardisation • Mind-sets + Interop /IOT Standards +
Wrappers
• Ma ufa turer s asi i sti t is to lock things down
FACEBOOK OF
MACHINES
Transition-4 :
From Isolated m/c to FB of m/c
Transition-5 :
From Traditional to Gamified command centre
Do 't desig a application, but alter a behaviour
Transition-6 :
From Pipe to Platform
Transition-7 :
From Data Streams to Revenue streams