April 18
6
C
over
S
tory
From Edge to Fog to Cloud –
IoT Computing with SMARC
By Martin Unverdorben,
Kontron
Cloud, Edge and Fog computing
is everywhere and everybody
is talking about it. In industrial
surroundings traditional server
approaches cannot provide the
required robustness for operation in
harsh industrial environments,
but how can modular systems based
on the tested SMARC Computer-on-
Modules provide an efficient solution.
The Industrial Internet of Things (IIoT)
is one of the most challenging application
spaces to design for, as there are pressures on
the developer from clients, management, and
the marketplace on price, performance, and
functionality. Creating the next generation of
intelligent industrial systems will require an
elegant juggling act with all three.
There is a trend in IT to bring more web
functionality out of the central server farms
and inject it into the parts of the infrastruc-
ture and the devices operating as close to the
user level as possible to reduce network traf-
fic demands, among other things. Data col-
lection and storage is already everywhere, in
our pockets and in our homes with personal
computers, smartphones, and smart home
assistants. It is also the case now in production
environments.
Cloud, Fog and Edge levels
Applying IoT technology to industrial systems
makes a lot of sense, because we can use and
manage the data involved to improve the pro-
duction process and performance factors such
as quality and cost. For example, in a chemi-
cal plant the process data is probably already
monitored, and some factors, like the tem-
perature and pH values of chemical reactions,
can be controlled. Why not store that in data
to analyze later? Another example is where
products parts are mounted or fixed together
with screws, where you store the torque val-
ues of each screw to have quality control of
all the devices that are mounted there. Peer-
ing through the Cloud Figure 1 looks at it from
a structural perspective. At the bottom are the
devices on the floor of the factory that con-
trol the production process directly, which is
the area called EDGE computing. On the level
above, there are the on-site server racks, which
manage the process flow, or control data and
monitor and maintain the data stored. The
layer on top is the CLOUD, representing all
off-premises functionality. The middle layer,
which usually already exists in some form in
legacy systems, gets new tasks because it must
connect to that top-level functionality level,
and is now called FOG computing.
So the Cloud is everything that is off the
premises, the Fog is the level that’s monitor-
ing and controlling, and used to be already
there before in the server room in the factory.
The lowest level, Edge computing controls the
floor and the automation control of motors,
sensors, and actuators. These three levels have
to function and work together, especially in
security and safety, with new tasks to be done
and new devices that need to be installed.
3 key functionalities of TSN
Another development coming up is TSN, time
sensitive networking. Started in 2012 as an
IEEE 802.1 working group, the TSN standard
defined a way of networking to ensure very low
transmission latency and high availability of all
participants. Originally intended to define a
network for real-time audio and video stream-
ing, this functionality is perfect to synchronize
control of devices on the factory floor.
There are three key functionalities involved. The
first is time synchronization, meaning all the
devices participating in real-time communi-
cation have a common understanding of time.
This can also be done with an internal clock,
but that can be cost-prohibitive. Using the IEEE
1588 time synchronization eliminates the need
for any extra clocks or any extra signals, as the
time information is distributed throughout the
whole network with this protocol.
The second key functionality is scheduling and
traffic shaping, where all devices adhere to the
same rules in processing and forwarding com-
munication packets. Already a known concept
from the telecom market, it means there are dif-
ferent slices for different traffic classes, allowing
you to give a certain packet a different class, or
a certain priority. The third is to ensure that all
devices comply with the same rules in reserving
bandwidths and time slots, possibly utilizing
more than one simultaneous pass to achieve
fault tolerance. That means TSN is a network
that ensures that all partners are talking on the
same time level, and also at a very high availabil-
ity and low transmission latency network.
