Edge Intelligence: The Central Cloud is Dead – Long Live the Edge Cloud!
approaches, adopted by companies such as
Google to minimize latency in response.
Others use scaled-down “out of the box”
data centers, co-located or situated near to
ISP nodes. This can be taken to its logical
conclusion, if such caching/hosting is
implemented at the base station level in
wireless networks. Finally this trend can be
extended to allow containerization at the
base station and thus enable the docking of
third party applications there.
In scenarios such as smart manufacturing,
with more and more robots, CNC machine
tools, and the like, generating massive real-
time data in the field, greater computing and
storage resources will be necessary. In such
situations a local cloud at the edge
represents a good choice. More edge servers
will be interconnected and provide pooled
and scalable resources. Several software
programs or services can be deployed on an
edge node simultaneously.
As billions of endpoint devices need to
connect, a critical component of future
Internet of Things systems “IoT gateway.”
New generation IoT gateways are opening
up huge opportunities to push processing
closer
to
the
edge,
improving
responsiveness and supporting new
operating models. The IoT gateway serves as
an important bridge between operations
and IT and also provides a cost-effective
business model by making use of lightweight
storage, networking, and computing that
can operate in industrial environments. By
adding IoT gateways, the current field
deployment could require no change in
order to run new applications such as
predictive maintenance on the gateway.
Software Evolution
Most of the IoT Edge Computing Node
technology, runs on flavors of the Linux or
Microsoft Windows (using the .net platform)
operating systems using various processor
architectures. An industry-wide trend is
emerging to package edge computing
capabilities into micro services and deploys
them within containers on IoT edge
computing nodes, as illustrated in Figure 1.
Containers are user space instances
executed in an operating system kernel
providing strong isolation between them.
The operating system kernel can provide
resources management between different
containers. Containers provide security
though isolation; they also serve as
deployment units that simplify lifecycle
Figure 4: Edge computing
- 20 -
September 2017