Scope and limitations of OPC-UA and TSN
OPC-UA and TSN are widely used buzz-words in the Industry 4.0 and industrial IoT debates. But what is behind these buzz-words, where are these technologies going and will they replace the fieldbus systems? Stephen Hayes of Beckhoff offers some guidance.
OPC-UA has established itself as a common denominator above the fieldbus level, and can be seen as the leading communication technology for Industry 4.0. Time Sensitive Networking (TSN), meanwhile, is anticipated to provide real-time capabilities for standard Ethernet. But can these technologies live up to the promise, and what do they mean for industrial networking?
The OPC Foundation emphasises that OPC-UA is not just a protocol but a service-oriented architecture. The development of OPC-UA (Unified Architecture) started in 2003. First prototype implementations were shown in 2006 and the specification was released in 2009. Making use of the client/server approach, OPC-UA provides data exchange mechanisms across different platforms. Data is transmitted via TCP/IP or HTTPS in either a special binary format or xml encoding (SOAP1.2).
If required, cyber security is established by standard security mechanisms such as TLS cryptographic protocols and authentication with certificates. OPC-UA specifies information models for actual and historical process data, alarms and also for more complex tasks such as programs. The object oriented information model approach also allows one to describe almost any system or data structure. The OPC-UA address space and services allow the clients to discover and access servers by browsing.
The richly featured architecture and especially the complexity of the protocols made the original OPC-UA communication fairly cumbersome compared to the fieldbus technologies. In order to improve the usability of OPC-UA in machine-to-machine applications and for vertical communication towards cloud based services, OPC-UA specifications have recently being enhanced by a publish/subscriber model and protocol. The publishing server sends data via UDP/IP multicast, and one or many subscribing clients can receive this data. For cloud communication OPC-UA Pub/Sub uses the Advanced Message Queuing Protocol (AMQP) or Message Queue Telemetry Transport (MQTT), which are relayed via broker services in the cloud.
OPC Foundation also plans to enhance the Pub/Sub protocols with features defined by the TSN task group of IEEE 802.1. This task group is handling a series of standardisation projects seeking to make Ethernet switching technology usable for time-critical applications. The projects take care of real-time topics such as improved synchronisation, transmission preemption, time scheduling and bandwidth reservation, but also cover redundancy and topology recognition. Several specifications have meanwhile been finalised, but lack a common solution with regards to configuration and administration. Other specifications are still in the making, and additional projects have been added recently.
TSN technologies will certainly enhance the application range for which Ethernet standards can be used. They are in line with previous Ethernet technology enhancements such as full-duplex communication, Quality of Service (QoS) or VLAN tagging, which are used by the IT world and the industrial Ethernet technologies alike. However, the over-optimistic cycle time data that floats around with regards to TSN make some people believe that the combination of OPC UA Pub/Sub and TSN will meet the real-time requirements of machine control applications and will thus make the fieldbus technologies obsolete, including the industrial Ethernet based fieldbus technologies. Interestingly enough these post-factual predictions are predominantly voiced by those who were not on the winning side of the fieldbus battles. But there are a number of technical facts as well as strategic reasons why this scenario is very unlikely, to say the least.
The background of the TSN task group is bridging (switching) of audio and video data (AVB). Consequently the protocols are optimised for larger amounts of data and are not ideal for devices with small amounts such as sensors, drives or I/O devices. The frame size requirements of Ethernet lead to an efficiency of below 10% in these applications. Furthermore, the forwarding methodology of IEEE 802.1 is ‘store and forward’ which produces a latency in the range of milliseconds for larger frame size and hop counts: for example, one single 500 byte frame over 20 hops generates more than 1ms of latency from sender to receiver. Aiming for real-time by optimising switch behaviour results in complex protocols and elaborate configuration, as can be seen in Profinet-IRT, which never really took off. Combining TSN with IP-based OPC-UA does not simplify the technologies either.
As of now, there are only few semantic models defined in OPC-UA. For instance, there is no common description on how to organise a linear movement or a rotation as defined in IEC 61800-7. Even sensors need some parameters and status information as defined in profiles which are not available for OPC-UA so far.
Neither OPC nor TSN are fieldbuses, nor were they intended to be ones. Key elements such as network management and device profiles are lacking. There are some attempts to map existing profiles, but this will not close that gap. Furthermore, the main strategic advantage of OPC is its neutrality for which the positioning above the fieldbus level is paramount. It is unlikely that OPC Foundation will jeopardise this asset.
OPC is a great technology which will become a pillar of Industry 4.0. With the Pub/Sub extension and the future integration of TSN features it will expand its application range. OPC can fill the gap of a standardised common language above the fieldbus level, but will not become a threat for the fieldbus systems itself – at least not for a high performance system such as EtherCAT which cannot be challenged by IP-based technologies. For example, due to the unique functional principle of EtherCAT called processing on the fly, EtherCAT entirely avoids the complexity of the internet protocols. It achieves not only an outstanding and exceptional performance which accelerates most systems and allows one to make use of the CPU power that modern industrial PCs provide, but at the same time is easier to implement and use even than the classical fieldbus systems, not mentioning other Ethernet based approaches.
EtherCAT is not based on IP-based protocols, but inherently supports IT technologies. Since EtherCAT and OPC-UA are complementary, not competing technologies, the EtherCAT Technology Group (ETG) was among the first fieldbus organisation to enter into a Memorandum of Understanding with the OPC Foundation – other organisations followed.
Industry 4.0 and IoT demand seamless communication through all layers and levels within the digital factory, as well as externally through cloud-based services and other Internet technologies. ETG and the OPC Foundation achieve these requirements with open interfaces between both of their technologies, rather than expanding their own technologies into the core areas of the other. TSN will enable the direct integration of EtherCAT segments into plant networks. Therefore ETG is actively contributing to the OPC Pub/Sub development as it has been contributing to the TSN development within IEEE from day one.
Other News from Beckhoff Automation Ltd
Latest news about Industrial Ethernet