Page 14 - EETEurope FlipBook February

14 EE|Times EUROPE

EMBEDDED
Industrial Computing Moves to the

Mission-Critical Edge

By Flavio Bonomi

cross a range of industries and
specifically in the industrial
automation vertical, there is
A broad agreement that the deploy-
ment of modern computing resources with
cloud-native models of software life-
cycle management will become ever-more
pervasive. Placing virtualized computing
resources nearer to where multiple streams
of data are created is well-established as the
path to address system latency, privacy, cost,
and resiliency challenges that a pure cloud
computing approach cannot address. This
paradigm shift was initiated at Cisco Systems
around 2010 under the label “fog computing”
and progressively morphed into what is now Figure 1: A distributed system of systems is intended to address the challenges of the
known as edge computing. many poorly connected, fragmented, and aging subsystems controlling today’s industrial
environments. Lynx has identified the evolution of the industrial operational architecture
MISSION-CRITICAL INDUSTRIAL (the architecture of the infrastructure on the industrial automation floor) as one of the
REQUIREMENTS most appropriate targets for the realization of the full mission-critical edge paradigm.
That said, the full potential of this transfor-
mation in both computing and data analytics
is far from being realized. The mission- of embedded computing (security, real-time • Distributed nodes that are remotely man-
critical requirements are much more strin- operation, and safe, deterministic behaviors), aged and software that is delivered and
gent than what the cloud-native paradigms into modern networked, virtualized, contain- orchestrated as virtual machines (VMs)
can deliver. This is especially true because erized life-cycle management and data- and and containers — the model of modern
mission-critical applications have specific intelligence-rich computing. cloud-native micro-services
requirements: Paired with new, more powerful, and scal-
• Heterogeneous hardware. Typical THE ROLE OF MISSION-CRITICAL EDGE able multicore platforms, a mission- critical
industrial automation settings have Without a fully manifested mission-critical edge computing approach can provide a
different architectures (x86, Arm), as well edge, we will not be able to address the many unified, uniform infrastructure going from the
as a variety of compute configurations on pain points characterizing the current indus- machine to the industrial floor and into the
the floor. trial electronic infrastructure. In particular, telco edge and cloud, thereby enabling a fun-
• Security. The security requirements and we will not be able to securely consolidate, damental decoupling between hardware and
their mitigations vary from device to orchestrate, and enrich with the fruits of data software. Applications, packaged as VMs and,
device and need to be handled carefully. analytics and artificial intelligence the many increasingly, as containers, can be life-
• Innovation. While some industrial poorly connected, fragmented, and aging cycle–managed and orchestrated across all
applications can continue with the legacy subsystems controlling today’s industrial the layers of this infrastructure.
paradigm of going unchanged for a decade environments.
or more, most of the industrial world now The broad architecture shown in Figure 1 INTEGRATION INTO TODAY’S
additionally requires modern data analyt- illustrates our vision for enabling this: FRAGMENTED INDUSTRIAL
ics and monitoring of applications in their • Distributed and interconnected, ENVIRONMENTS
installations. mixed-criticality–capable, virtualized Many of the poorly connected, fragmented,
• Data privacy. As in other areas of IT, data multicore computing nodes (system of and aging subsystems controlling today’s
permission management is increasingly systems) physical environments can be effectively
complex within connected machines and • Networking support that includes tradi- and securely consolidated, orchestrated, and
needs to be managed right from the origi- tional IT communications (e.g., Ethernet, enriched with the fruits of data analytics and
nation of the data. Wi-Fi) but also deterministic legacy field artificial intelligence.

• Real-time determinism. The real-time buses, moving toward IEEE time-sensitive Figure 2 shows how the infrastructure
determinism provided by controllers networking (TSN) and public and private would look when the mission-critical edge
remains critical to the safety and security 4G/5G, also moving toward determinism is deployed and embedded into the oper-
of the operation. • Support for data distribution within and ational technologies area of the factory.
For these reasons, the market is seeking across nodes, based on standard middle- There is a distributed set of nodes —
what Lynx Software Technologies calls the ware (OPC UA, MQTT, DDS, and more), some very close to the plant, some far
mission-critical edge. This concept is born out that also strives toward determinism (e.g., away. Effectively, this is like a distributed
of the incorporation of requirements typical OPC UA over TSN) data center, but it contains a far more

FEBRUARY 2021 | www.eetimes.eu