1. ALMANAC Internet of Things for Smart
Cities
Dario Bonino, Delgado Alizo M.T., Alapetite A., Gilbert T., Axling M., Udsen
H., Carvajal Soto J.A., Spirito M.
2. SMART CITY: LAYERS BEYOND TECHNOLOGY
Life Services
Data / Information
Smart Infrastructure
Basic Infrastructure
3. ONE LAYER AT TIME?
Current approaches mostly focus on a subset of
the smart city layers
Data / Information
Smart Infrastructure
4. RELATED WORKS
EC-funded projects
FIWARE – generic enablers for smart city processes
Open IoT
Smart Santander
Urban Water
Private company initiatives
IBM
HP
BOSH
CISCO
Academic Research
KM4City
5. ADDRESS SMART CITY AT ALL LAYERS
Novelty of the ALMANAC approach: try to
address all layers by design
1
Smart Cities as a network of stakeholders,
both private and public
2
Federation as key to support legal /
administrative agreements between parties
3 Service distribution and Thing independence
6. 30’000 FEET VIEW
Federated cities exchanging
users and data
DNS-based platform naming
Geo-DNS redirection
(under testing)
SSO for users and devices based
on city-level agreements
7. CITY-LEVEL VIEW
Federation of different entities in a single city
Each entity may run one or more “Platform Instances”
Not all entities must be part of the same federation
Waste Collection
ServiceWater Distribution
Municipality
8. SCENARIOS
Waste collection Water Management
Citizen-centric Applications
Waste bin monitoring
Waste issue collection and tracing
Route optimization / planning
Tariff / Cost optimization
Water metering
Leakage detection
Disaster management (naïve)
Recycling guide, Waste Bin locator, Crowdsourcing for waste disposal
Bicycle reservation, Bicycle locator
Issue reporting
9. INGREDIENTS
Distributed approach based on “Platform Instances”
Per quarter / city / entity
Peer to Peer Federation of “Platform instances”
Shortens the path to data and resources
Different “realms” to account for different data
access / visibility
Reflecting agreements between involved entities
Micro services approach
Scale-up to billions of data
Field-access
Access and “abstract” field-level data exploiting semantics
10. PLATFORM VIEW
Almanac Cloud APIs
Virtualization Layer
Data Management Layer
Resource Adaptation Layer
REST & MQTT APIs for applications
Peer2Peer platform communication,
Service coordination,
Federated Identity Management
Complex Event processing and Data fusion,
Observation storage, Resource catalog,
Semantic assets
Device / Object access and abstraction
Data validation,
Semantic annotation
IoT-A – compliant platform design
11. ALMANAC CLOUD API (1/2)
Smart City Resource Library Services API
Query for/look up IoT Resources and Things based on
metadata (Semantic Library Services).
Historical Data API
Access to stored observations in data streams (time series
data) from sensors or data fusion queries.
Live Data API
Subscription to streams from sensors or data fusion
queries (e.g. using web sockets).
12. ALMANAC CLOUD API (2/2)
Data Fusion Services API
Enables applications to generate new data (data
streams) by defining complex event processing queries
Provisioning & Management API
Resource provisioning
Users, Groups and Privileges (RBAC)
13. VIRTUALIZATION LAYER
Linksmart
P2P platform interconnection
MQTT data stream forwarding
PEP implementation
Virtualization Layer Core
API coordination / redirection
Swagger-based descriptions
Platform component coordination (e.g., through Node RED)
Format adaptation
Federated Identity and access management
SSO and Access Control
14. DATA MANAGEMENT LAYER (1/2)
Resource Catalogue
IoTEntities (OGC Things) descriptions
Search & Retrieve (OGC Sensor Things API)
Local Discovery (UPNP)
Latest data (retrieved either via REST or UPNP)
Storage Manager
Historic data series
Exploiting cloud-based deployment (Azure) and NoSQL
(MongoDB)
15. DATA MANAGEMENT LAYER (2/2)
Data Fusion Manager
Complex Event Processing
Data Fusion Language,
for defining block-based processes through REST APIs
Semantic Representation framework
Entity representation through ontologies
Smart city ontology (exploits a branch of DogOnt)
Waste bin ontology
Many linked models (e.g., GeoSPARQL, Places, GoodRelations, etc.)
Query and retrieval
Direct SPARQL endpoint
Predefined templates for easier access
16. RESOURCE ABSTRACTION LAYER
Smart City Resource Adaptation Layer (SCRAL)
Heterogeneous devices and protocols
Generates OGC SensorThings compliant Observations
Asynchronous delivery through MQTT (QoS 0)
Device-level PEP
REST APIs for “smart devices”
To push data inside the platform
17. IMPLEMENTATION
Micro-services pattern
Services distributed on multiple locations,
hardware, with different OS and computational
power
Service to platform association by means of DNS
naming
Different technologies and solutions
Java, C#, Python, MongoDB, UPNP, etc.
18. EXPERIMENTAL SETTING
Scalability / Feasibility
3 different PIs in 1 Federation
Over 60’000 synthetic sensors connected to a single
SCRAL Instance
29k waste bins
24k water meters
1k waste bins from Smart Bin
Selected sensors from Smart Santander
Sensors from the TiLab M2M server
19. ON THE FIELD DEPLOYMENT
DEPLOYED
8 Underground Ecological Islands in
Turin
2 Card readers (RFID) for the non-
recyclable fraction
1 Weight sensor on the collection truck
Cellular-based capillary network for
continuous data collection
FEATURES ENABLED
Fill-level monitoring
Dynamic collection routes
Issue reporting and management
20. LESSONS LEARNED / RESULTS
Technology issues are tough but can be addressed
Big data
Distributed computing
Cloud services
Administrative, Legal and Privacy issues are typically
underestimated
Complex, hard to settle
Regional policies are often incompatible and difficult to
generalize
Entities are “resilient” to data sharing,
difficult to set up efficient exchanges
21. QUESTIONS? Internet of Things for Smart
Cities
Dario Bonino, bonino@ismb.it
The project is co-funded by the European Commission under grant
agreement 609081.
This presentation reflects solely the views of its authors. The
European Commission is not liable for any use that may be made of
the information contained therein