Towards Configuration Technologies for IoT Gateways

Workshop on Configuration 2016, Toulouse, France
Institute for Software Technology
1
Towards Configuration Technologies
for IoT Gateways
A. Felfernig1, S. Polat Erdeniz1, P. Azzoni, M. Jeran1,
A. Akcay1, and C. Doukas
1Graz University of Technology, Austria
2Eurotech Group, Italy
3Create-Net, Italy
alexander.felfernig@ist.tugraz.at
Workshop on Configuration
Novi Sad, 25.-26. Sep. 2014
FinRec‘16
Bari, Italy, June 16th 2016
Workshop on Configuration 2016

2
Overview
•AGILE Project
•Example IoT Application Scenarios
•Ramp-up configuration
•Runtime configuration
•Research Issues
•Conclusions

3
Horizon 2020 Project AGILE
• AGILE: an Adaptive & Modular Gateway for the IoT
• Duration: 2016 – 2018
• 17 Partners:
• Goal of TU Graz: recommendation & configuration for IoT

4
Horizon 2020 Project AGILE
• Internet of Things (IoT): paradigm that envisions a
networked infrastructure enabling different devices
(things) to be interconnected at anyplace and anytime.
• Gateways: bridge devices to corresponding applications
on the basis of different protocols (e.g., temperature
sensor with weather app on Raspberry Pi).
• Ramp-up Configuration: initial determination of
gateway infrastructure (devices and applications).
• Runtime Configuration: reconfiguration of devices (e.g.,
sensors, drivers, and app settings).

5
IoT Application Scenario:
Smart Homes
• Sensors: temperature,
body temperature,
heartrate, air pressure,
incoming light, motions
• Apps: personal trainer,
burglary protection,
automated watering, fall
detection, light
management,
personalized receipes

6
Animal Monitoring
• Sensors: body
temperature, heartrate,
GPS data, ground photo-
graphs from drones, air
temperature data.
• Apps: disease detection, detection of forest fires,
observation of herd movements, fighting of poachers,
search for individual animals, remote treatments, “lost &
found” apps, mobile camera management, statistics (e.g.,
population sizes in herds).

7
Enhanced Retail Services
• Sensors: own location,
item locations, items in
shopping cart, locations
of experts.
• Apps: personalized recommendations & special offers,
personalized content displays, gamification-based
shopping experiences, competitions, personalized item
information, retrieval of experts.

8
Ramp-Up Configuration
• Initial determination of gateway infrastructure.
• Smart Homes: given information about rooms, levels,
etc., and the preferences of users, propose a
corresponding configuration per room (sensors, etc.).
• Animal Monitoring: given the topology of the area incl.
the types/amounts of animals, derive a configure-tion of
devices (e.g., drones, measurement units) needed.
• Enhanced Retail Services: given the size of the shop
and the different types of items, provide a configuration
that includes the needed devices and their positioning
(e.g., monitors and sensors).

9
Runtime Configuration & Recommendation
• Recommendation basis: gateway profiles (e.g., apps and drivers)
of other users.
• Configuration & reconfiguration have to take into account
dependencies between items (e.g., driver x requires driver y).
• Local configuration: changes of drivers/protocols, etc.

10
Runtime Configuration & Recommendation
• Recommendation basis: app purchasing behavior of other users.
• Configuration & reconfiguration have to take into account
dependencies between between apps.
• Local configuration: changes of drivers/protocols, etc.

11
Runtime Configuration & Recommendation:
Summary
• Apps: given the set of currently installed apps and
sensors, recommend additional apps and devices. This
includes a reconfiguration of the current gateway settings.
• Collaborative Recommendation of Apps. Based on
gateway profile repositories and app repositories.
• Content-based Recommendation of Workflows. Based
on workflow repositories such as nodered.org (workflows
published by other gateways).
• Collaborative Recommendation of Sensors. Based on
gateway profile repositories and app repositories.

12
Current Research Issues
• Knowledge representation and reasoning (CSP, ASP,
…) that supports the mentioned ramp-up scenarios.
• Possibilities to improve test case generation and
management for complex configuration knowledge bases
(on the basis of existing SWE approaches).
• Increasing the performance of personalized
diagnosis and configuration search (e.g., anytime and
direct diagnosis).
• Knowledge compression techniques for configuration
processes on the gateway (runtime configuration).

13
Performance of CSP Solver
(Preliminary Test)
• Comparison of CHOCO constraint solver on Raspberry
Pi 2 (Snappy) and Lenovo E50 with LinuxMint.
• Detailed analysis in scope of future work.

14
Performance of AnyTime Diagnosis
• AnyTime diagnosis (with limited minimality): worst case
estimates for the number of needed consistency checks
depending on the granularity m.
• Detailed analysis in scope of future work.

15
Conclusions
• Many existing IoT scenarios, for example, smarthomes,
animal monitoring, and enhanced retail services.
• Configuration plays a major role in IoT scenarios.
• Often combined with recommendation (e.g.,
recommendation of apps and gateway reconfiguration).
• Challenges in terms of knowledge representations
(configurations take into account time and space).
• Challenges in terms of algorithm efficiency on the
gateway (limited computational resources).

16
Thank You!

Towards Configuration Technologies for IoT Gateways

Empfohlen

Empfohlen

Weitere ähnliche Inhalte

Ähnlich wie Towards Configuration Technologies for IoT Gateways

Ähnlich wie Towards Configuration Technologies for IoT Gateways (20)

Mehr von AGILE IoT

Mehr von AGILE IoT (20)

Kürzlich hochgeladen

Kürzlich hochgeladen (20)

Towards Configuration Technologies for IoT Gateways