Maximizing Efficiency and Profitability with OnePlan’s Professional Service A...
Confluent Cloud transforms ASPI's digital transformation
1. Confluent Cloud – Data in Motion
06/10/2022
1
Confluent Cloud inside
the Digital
Transformation of
Autostrade per l’Italia
Confluent Cloud – Data in Motion
06/10/2022
2. Confluent Cloud – Data in Motion
06/10/2022
1. Company Presentation
2. ASPI Digital Transformation
3. The use of Kafka in ASPI
4. Confluent Cloud Migration
5. Next Steps
AGENDA
3. Confluent Cloud – Data in Motion
06/10/2022
3
2.855 km
of motorway in Italy
6,1 Billion €
+110% of investiments
in 2020/24 vs 2015/19
nearly 5.300
employees
204
Service Areas
1.907
Variable-message signs
along the network
4.991
cameras
3 Billion
€
of revenues Autostrade per l'Italia is one of Europe's leading
concessionaries for the construction and management
of toll motorways, with around 3,000 km of network
managed in Italy.
The Autostrade per l'Italia group includes five other
motorway concession companies and seven
companies that provide ancillary services to the core
motorway business and play a key role in
implementing the Company's transformation plan.
• Società Italiana
Traforo del Montebianco
• Raccordo Autostradale Valle d’Aosta
• Tangenziale di Napoli
• Autostrade Meridionali
• Società Autostradale Tirrenica
• Movyon
• Tecne
• Amplia Infrastructures
• Free-to-Experience
• Elgea
• Admoving
• Esse di Esse
• Giove clear
THE GROUP
THE COMPANY
2,5 Billion €
+60% of maint. costs
in 2020/24 vs 2015/19
4. Confluent Cloud – Data in Motion
06/10/2022
«Next To Digital» INITIATIVE
The Next To Digital initiative is included in the wide
Transformation Plan of Autostrade per l’Italia.
This initiative is the fundamental lever to guide and promote a
deep change in ASPI at multiple levels: new ways of working,
new technological capabilities, new skills to develop, new ways
of addressing all stakeholders (internal and external) and, finally, a
significant cultural change.
It concerns the digitalization of processes and services,
transformation of the road network into a "smart" infrastructure
through the most advanced technologies.
NtD initiative address these aims:
• Internal efficiency
• Civil infrastructure management and security
• Customer experience
200 Million €
of investment by 2024
+ 30 agile room
to develop End-to-End
innovative projects
5. Confluent Cloud – Data in Motion
06/10/2022
«Next To Digital» PROJECTS
Predictive models of traffic flows for planning construction sites
to minimize the impact on the road network (estimation of any
inconvenience caused and calculation of tariff discounts)
Real-time monitoring and alerting systems in relation to traffic
disruptions (e.g. road accidents) or road congestion on the network
due to:
- Proactive and dynamic management of lanes and speed limits
- Preventive deployment of traffic/rescue teams
New services for users: Recommended times for departure;
advance calculation of the toll; recommended restaurants,
attractions and places of interest along the route
Additional services in the service area (e.g. parking, "skip the line"
table reservation) and on the network (e.g. advance toll payment)
Digital communication channel between the customer and
Autostrade for distress alerts requests, complaints, assistance and
automatic processing of requests via Virtual Agent.
6. Confluent Cloud – Data in Motion
06/10/2022
Why Kafka?
• Become a Data Driven company
• Create a unique data layer to support Event Driven
architecture
• Eliminate Data Silos
• Provide Data Democratization through REST API
• Provide Standardization through JSON, AVRO
• Data Security with Zero Trust approach
Confluent Kafka has been chosen to accomplish Autostrade per l’Italia’s goals:
7. Confluent Cloud – Data in Motion
06/10/2022
Business Cases and Architecture
8. Confluent Cloud – Data in Motion
06/10/2022
Online Monitoring Traffic
GOAL
Monitor and analyze transit on tracks collected at toll booth in near real time.
HOW
RESULTS
Change-Data-Capture pattern of Confluent Cloud with 2 sync connectors.
• RAW data converted to JSON
formats for easier usability with
REST API
• Freshness of information for
end users
• Decoupling systems for less
overhead on DB2
9. Confluent Cloud – Data in Motion
06/10/2022
Electronic Toll enrichment
GOAL
Implementation of an on-cloud application that allows the management
of every aspect of the rectification traffic life cycle for users and special
users.
HOW
RESULTS
Raw Data is normalized and democratized through Kstream Java App
With Confluent Kafka, through the KStream we are able to deal with
data that was previously on silos to have a gain on data governance
on the outgoing DBMS.
10. Confluent Cloud – Data in Motion
06/10/2022
Integration Layer "CEM Impianti"
GOAL
Standardize the monitoring data and render them accessible via front end for reporting / monitoring activities.
HOW
RESULTS
Unique Integration Layer to collect data from OT Plants based on 3 steps:
• Mapping of monitoring data
• Data definition and Normalization
• Front-End for reporting and analytics
• Greater usability of data
• Reduced interventions time
• Best use of resources
• Reduction of low added value
activities
11. Confluent Cloud – Data in Motion
06/10/2022
PISTA 2.0
Most complex project based on Confluent Cloud (in evolution…)
GOAL
Refactoring of the message production and dispatching system from all the toll booths on the highway to a single Data Layer.
HOW
RESULTS
Provide a formal definition of the data using AVRO Schema, useful for:
• Manage information as a resource
• Create a process aimed at managing the data
• Data preparation and enrichment
• No use of DB for first ingestion of message acquisition
• Event Streaming – Ready
• Provide data detectability, monitoring and security
throughout its life cycle
12. Confluent Cloud – Data in Motion
06/10/2022
PISTA 2.0 Architecture
13. Confluent Cloud – Data in Motion
06/10/2022
13
13
Confluent Platform migration
to Confluent Cloud
14. Confluent Cloud – Data in Motion
06/10/2022
Migration to Confluent Cloud
SCENARIO
The migration from Confluent Platform to Confluent Cloud involved 15 Applications, mostly
based on collecting data from tracks and plants.
PROJECT
USE CASES
Autostrade per l’Italia started a multi-year modernization project that relies on the
transition from on-premise systems to Cloud platforms for increasing speed to
accomplish business requests and reduce low value-added activities.
• Pista 2.0 - First acquisition of transits and plate tracking on the toll booth
• Plants - Spooling of messages from suburbs to center
• TUTOR - collecting events
• Infoviability - collecting data from sensors on track
• Xmonit - Realtime monitoring and integration of plants
• FreeToX – Cashback system for customers
• Service Area – Smart services on the services area on the highway
15. Confluent Cloud – Data in Motion
06/10/2022
Confluent Platform
• 3 different environments on-premises
• Ansible playbooks for configurations
• Custom connectors
• Manage infrastructure for updates and develop new features
• Not scalable
• Storage tiering needed
• Managing Load Balancing for Kafka Brokers
ENVIRONMENTS
PROS
CONS
SVIL TEST PROD
16. Confluent Cloud – Data in Motion
06/10/2022
Confluent Cloud Architecture
ENVIRONMENTS
• Prod and noProd environments for data segregation
• Teampass as a Password Manager for Client API Keys & Secrets
• Zabbix as ASPI main Monitoring Platform
PROS
CONS
• 99.99% SLA
• Autoscaling
• No patching or planned maintenance
• Infinite Storage
• Easily connected via AWS Transit Gateway
• Users ACL: access to Control Center for Data Lineage view
• Kafka Connect onprem for custom use cases
17. Confluent Cloud – Data in Motion
06/10/2022
Next Steps
TRAFFIC PROJECT
IoT/MQTT
Refactoring of the functional perimeter, based on more than 70 applications, linked to the
collection and management of payments and transit data on the track. Main problems are:
• Highly interconnected area with strong dependencies between functional levels
• Highly standardized architectures based on legacy technologies and integration models (eg.: DB2,
db-to-db views).
• No API
• No Data Democratization
• No SecDevOps approach
• Solution: Confluent Cloud as single layer for Data acquisition & preparation.
Provide an enterprise platform for MQTT to manage messages from IoT devices and lay the
foundation for Event Driven Architecture.
• Main products identified: EMQX, HiveMQ, VerneMQ.
18. Confluent Cloud – Data in Motion
06/10/2022
18
THANK YOU
Confluent Cloud – Data in Motion
06/10/2022