11.30 - Kiran Pande - BUILDINGS V1

Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

Fast Track IOT & Analytics in

Cloud
- Kiran Pande

1
This is what we will cover

01 Airport Analytics – A cursory look at

02 A typical messaging day at Airport

03 IOT & Analytics is Cloud Bound – Some basics

04 A typical implementation

05 Problems with traditional IOT & Analytics

06 What are options?

2
A cursory look at Airport Analytics

Improved Operational Planning Improved Staff Morale


• Telemetry of processes • Efficient Operations Lower Employee Stress
• Evaluation of process bottlenecks • Measurable Goals
• Data heatmaps for process problems. • Improved Customer Experience

Performance Monitoring Better Security


• Data insights to Further Planning • Reduce False Positive Alarms
• Evaluation of Plan Execution (How?) • Improve Signalling
• Airport’s capability statement • Continuous Training

Reduce Wait Time Improved Baggage Performance


• Predict Airport in-out Flow • Manage inbound-outbound luggage
• Monitor - Curb to Gate – 30 mins • Reduce DieBacks n Chokes
• Improved Customer Experience • Sort Efficiently (eco, premium, first)

IMPROVED CUSTOMER SATISFACTION


3
A “Typical Messaging Day” at Airport

(schedules, estimated, expected, gate


open, gate closed, boarding, taxied, Flight Messages - 0.1m to 0.2m messages per day
return to stand, Airborne)

(At check-in, self check-in, at-gate,


PAX Messages - 0.5m to 0.8m messages per day
boarding, boarded, return to terminal)

(transfer, auth to load, remove, re-flight,


Terminating, screening, identification, Baggage Messages - 7m to 9m messages per day
Bag delivery, Loaded, Missed, Problem,
Manual Load etc)

(At check-in, self check-in, at-gate,


Service Messages - 24m to 27m messages per day
boarding, boarded, return to terminal)

4
IOT & Analytics is cloud bound

Network of Things has always been there Internet of Things is its extension

Connected Devices were operational use Now you locate, monitor and manage,
Analyse Connected devices – from cloud

First things first

Add on devices to Add the business


convert from information can transformation
analog signals to then be moved capabilities (remote
digital signals onto a network access, predictive
maintenance)

Data generated by the connected machine needs to be transformed into something


that adds intelligence

5
Traditional Implementation

Queuing Layer

Lake Layer

Orchestration Layer

Warehouse Layer

Access Layer

Kafka, Kinesis, RAW, Curated, Spark, Glue, Data Staging, Transformation, Dashboards, Reports,
6
Service Buses Aggregated, Tokenised Factory/Polybase Production Apps, Alerts, Notifications
Problems with typical IOT + Analytics

• Loss in Early Data Availability


• Time to Market/Access
• Longer Projects
• Loss of speed to react
• Costs

7
How to Speed up and save costs in Simple Architecture
30 Seconds end to end

Semantic Calculations,
Streaming Aggregations
Layer

Push
Datasets
Event Stream
Hub Analytics

Visualization
Storage and Reporting

PowerBI PowerBI
Report
Azure
Azure Server
Blob
Data Lake
Store

££££ Savers
• No Expensive DB Server
RAW Data • No Expensive Compute
3 Months MVP Access • Reduced data movement
• Reduced Org SW Footprint
6 Months 1st Product 8
Serverless Computing fast tracks IOT Analytics
Equipment Alarms

Satellite/Freq Tower Comms


Machine
Learning

Airfield Radio Msgs


Raw Power BI
Filter Noise in Data Dashboard
Data Lake
Robots Comms Curated

Movement

• Risk Models
• Predictive Maintenance
Machines Heartbeat
• Disruption management

9
Thank You & QA

10

You might also like