Oniro is a distributed operating system for mobile/IoT devices envisioned by Huawei and other partners. In such a scenario, our project aims at designing a middleware for dispatching workloads on a distributed architecture composed of several IoT, mobile and edge devices. The project focuses on intelligent orchestration of the workload in a dynamically changing ecosystem.
SFScon21 - Antonio Miele - EDDIE - Enabling Distributed intelligence in Oniro
1. Titolo presentazione
sottotitolo
Milano, XX mese 20XX
Enabling Distributed Intelligence in Oniro
Francesco Amigoni
AIRLab
Politecnico di Milano
Antonio Miele
NECSTLab
Politecnico di Milano
2. Francesco Amigoni and Antonio Miele
Working context: Oniro
Oniro is an Eclipse Foundation project focused on the
development of a distributed open source operating system
for consumer devices, regardless of the brand, model, make
• Portable across mobile and IoT ecosystem
• Vendor-neutral OS
• Enabling a global view on the network of devices
3. Francesco Amigoni and Antonio Miele
EDDIE project
In September 2021 we started a project called EDDIE “Enabling
Distributed Intelligence in Oniro” with the general goal of
developing a middleware system that is
“capable of giving to the applications a unified global view on the
ecosystem, so that the various devices are not seen any more as
individual entities with specific resources but as a unified pool of
virtual resources that can be exploited for an efficient
execution of any application required by the user and/or
interacting with the environment”
4. Francesco Amigoni and Antonio Miele
Examples of scenarios: from simple…
Use the mobile phone to control any smart appliance
• To open a door with a smart lock
• To order a coffee at the vending machine
• To start home appliances
5. Francesco Amigoni and Antonio Miele
…to more complex ones
• Use the phone camera for
authentication to open a door
with a smart lock
• Make the music to follow a
person moving in the home
• Customize lights in the home for
the specific family member
Need for complex orchestration
and dynamic reconfiguration!
6. Francesco Amigoni and Antonio Miele
…to more complex ones
• Use the phone camera for
authentication to open a door
with a smart lock
• Make the music to follow a
person moving in the home
• Customize lights in the home for
the specific family member
Need for complex orchestration
and dynamic reconfiguration!
7. Francesco Amigoni and Antonio Miele
The same example – system point of view
My mobile
phone
Smart TV in the
living room
Edge device on
the entrance
DEVICES
RESOURCES
APPLICATIONS
LOCATION
8. Francesco Amigoni and Antonio Miele
The same example – An application enters the system
My mobile
phone
Smart TV in the
living room
Edge device on
the entrance
Door
open
appl
DEVICES
RESOURCES
APPLICATIONS
?
?
LOCATION
9. Francesco Amigoni and Antonio Miele
The same example – Classical approach
My mobile
phone
Smart TV in the
living room
Edge device on
the entrance
Door
open
appl
DEVICES
RESOURCES
APPLICATIONS
LOCATION
10. Francesco Amigoni and Antonio Miele
The same example – Our envisioned solution
My mobile
phone
Smart TV in the
living room
Edge device on
the entrance
DEVICES
RESOURCES
APPLICATIONS
Door
open
appl
LOCATION
11. Francesco Amigoni and Antonio Miele
The same example – Our envisioned solution
My mobile
phone
Smart TV in the
living room
Edge device on
the entrance
?
?
Door
open
appl
DEVICES
RESOURCES
APPLICATIONS
virtual resources
LOCATION
12. Francesco Amigoni and Antonio Miele
The same example – Our envisioned solution
My mobile
phone
Smart TV in the
living room
Edge device on
the entrance
Door
open
appl
DEVICES
RESOURCES
APPLICATIONS
virtual resources
LOCATION
13. Francesco Amigoni and Antonio Miele
The proposed EDDIE middleware
We think of an open and dynamic
system organized on 4 different
levels
The goal of the project is to
design the EDDIE middleware
14. Francesco Amigoni and Antonio Miele
The proposed EDDIE middleware
Basic middleware infrastructure
• Communication between devices
• Device registration and
discovery
• Message/data exchange
• Distributed resource virtualization
• Make resource accessible from
remote devices
• Distributed dispatching and
scheduling
• Start an application on a remote
device
15. Francesco Amigoni and Antonio Miele
The proposed EDDIE middleware
Intelligent device orchestration
and application execution
• Define descriptors (ontologies!) for
applications, devices, and
planimetry. Describe:
• Application requirements
• Device resources and location
• Design a smart engine to identify at
runtime the optimal execution
configuration
16. Francesco Amigoni and Antonio Miele
The proposed EDDIE middleware
• Applications: requirements
• Required CPU and bandwidth,
required peripherals, which
location to interact with, …
• Devices: resources and location
• CPU and bandwidth
capabilities, connected
peripherals, location in the
planimetry…
• Each aspect can be in turn
parametrized
17. Francesco Amigoni and Antonio Miele
… To be continued!
EDDIE is expected to enable distributed intelligence in Oniro
First results are expected in 2022: stay tuned!
Contacts:
• francesco.amigoni@polimi.it
• antonio.miele@polimi.it