Edge Intelligence: The Central Cloud is Dead – Long Live the Edge Cloud!
In the world of software-defined networking
(SDN), the intelligence resides in the SDN
controller that informs one or multiple SDN
switches how to route packets (Figure 3-27).
The SDN switches are programmable
switches that interface via a protocol with
the SDN controller. Monitoring of the data
flows is also supported so that the SDN
controller can be informed about the traffic
the SDN switches are handling.
connecting end devices across the range of
access networks from Wi-Fi and LTE towards
NB-IoT, unlicensed spectrum LTE (LTE-U) and
ubiquitous satellite connections.
U SE C ASES AND R EQUIREMENTS FOR
E DGE I NTELLIGENCE
Factory Productivity Improvement
For factory operators, it is important to
improve productivity to maximize profits
and minimize waste. Big data analysis
realized through a combination of edge
computing and cloud computing can
contribute
substantially
to
factory
productivity improvement. Edge computing
with intelligence converts data collected
from the shop floor into data which can be
analyzed by cloud computing. The analysis
result leads to reduction in number of
products identified to be defective by
mistake and also enables predictive
maintenance, which can reduce factory
down time.
The flexibility of the software-defined
networks is that initial switching algorithms
or policies can be provided to the switching
layer
and,
based
on
rules/policies/algorithms provided by the
edge nodes, the master switching entity can
adapt and improve the edge switching
intelligence. The SDN switches can receive
algorithms to process the data, and based on
the traffic that is monitored can improve the
algorithm and upload it to the SDN
controller. Thus, SDN provides highly
efficient and low-cost automatic Operation
and Maintenance (O&M) and realizes policy
collaboration, convergence of networks and
security.
Connected City Lighting
Network Slicing is a technology that allows
Service Providers to build their own edge or
core network using core network software
components.
While bringing safety, convenience, and
ambience to urban citizens, city lights
consume a vast amount of energy and
operational costs for municipalities. To
address these issues, connected city lighting
is being developed to optimize lighting
coverage and resource usage. Firstly,
integrated smart lighting policies ensure
efficiency by controlling lights according to
time, ambient light conditions, and weather.
Secondly, online monitoring of each light’s
status in real time can automatically alert
the maintenance through edge computing
node once a malfunction occurs. The
The 5G Core Network standardized by 3GPP
takes advantage of the SDN as well as
network slicing technologies, allowing for
isolated dedicated networks to be built and
managed. The networks can then adapt
based on triggers coming from the
application domains, e.g. alarms from
detected security threats or other events.
Due to the softwarization of the 5G Core
Network, there is a uniform approach for
IIC Journal of Innovation
- 23 -