wilamowski-b-m-irwin-j-d-industrial-communication-systems-2011

67-4 Industrial Communication Systems
method nowadays is using electronic chip cards (a.k.a. smart cards), trusted platform modules (TPM),
or other chips that contain the key and algorithms to use the key. With this, the problem of identity and
trust is—again—transferred further: to the administration of manufacturing and shipping the devices.
Each one must get its own personal chip that defines its identity and helps with security.
The second big challenge for M2M communication is network management. As networks are getting
larger, it is required that traditional network administration tasks are automated. Assigning addresses
and baud rates is already past; most of these things happen in a plug-and-participate way. The next step
is plug-and-work, where also the network higher layers and the application are configured automatically.
The key ingredients for this are
• Service discovery and lookup tables
• Self-description of services and capabilities
• Application profiles
Naturally, not every application can be configured automatically without human involvement, but
learning network management tools can ease that job dramatically.
67.6 Scalability in Hardware and Software
The trend for “total integration,” i.e., the as-seamless-as-possible connection of different applications
and networks, opens up the reason for the very existence of different networks. Their different
features that made them suitable for specific application areas might be a barrier for integration.
One of them is size. Size in this context partly means the physical size, but more the code size of protocol
stacks, the silicon size for embedded functionality, and ultimately the price tag. Combining
low-cost sensor networks with powerful multimedia or real-time networks does not work out-ofthe-box.
What is needed is a scalable architecture where—in terms of hardware, protocols, and software—very
different nodes can interoperate, although they might only share a small set of common
services [PAL04].
One first step toward such a scalable architecture is the introduction of 6LoPAN, a downsized version 6
IP over IEEE 802.15.4 (the popular wireless transport, also used by ZigBee and other technologies).
Other wireless technologies like ISA-100.11a and WirelessHART show similar potential. But the idea
should not be limited to wireless technologies. The concept of scalable architectures must be technology
agnostic, ensuring that media access, addressing, management, and service discovery follow some
minimum core requirements that allow for connecting a $0.5 node to a $500 node in order to share some
functionality.
The Instrumentation, Systems, and Automation Society (ISA) defines six classes of industrial
communication:
• Class 0: Emergency
• Class 1: Closed-loop regulatory control
• Class 2: Closed-loop supervisory control
• Class 3: Open-loop control (human in the loop)
• Class 4: Alerting
• Class 5: Logging, downloading
Class 0 falls into the category safety, the other classes are either control (1 and 2), monitoring (4 and 5),
or something in between. Depending on its class, a certain service or application assumes certain
QoS, and operating applications with different classes over one and the same transport is not easy. ISA-
100.11a suggests using IEEE 802.15.4 as physical and TCP/UDP/IPv6 as network and transport layers.
For Class 0, however, even wired networks are often not sufficient. It might still be a long way until a real
universal and scalable network is available.
© 2011 by Taylor and Francis Group, LLC