Industrial Networks Connecting Controllers via OPC

+ MDT), whereas MTBF is the “mean time between failure” and MDT the “mean down time” [2]. However, in this thesis, availability is used in a more general manner, since the basis for mathematical operations is not available.

6.2 Components and Terms • Reliability means the probability of a device remaining failure free during a specified time interval, e.g. the maintenance interval: R = e λt

• Redundancy is the implementation of extra components in addition to the ones needed for normal operation. Thus, redundancy normally increases reliability and availability. 

2.2 OPC OPC, originally short for “OLE for Process Control”, is an open, standardized software communication interface specification launched in 1996 by a task force of different automation companies, later forming the OPC Foundation. As the former name indicates, OPC is an adaption of Microsoft’s Object Linking and Embedding OLE1 to the process control business, which used to be highly proprietary at that point of time. Thus it was almost impossible to efficiently combine products of different vendors. By providing so-called OPC servers with their devices, buses and software, vendors open their products to any OPC compliant client able to connect to the server for data exchange. Usually, an OPC server can handle several clients at once, while these clients—e.g. visualization or calculation applications—can connect to different servers in order to obtain their needed information. Figure 2.2: Classical OPC configuration Over the years, the OPC Foundation has been adding eight additional speci- fications to the original one, therefore the name OPC was freed from its original meaning and is now used as an umbrella term [3]. Some important specifications are quickly explained in the following: 

• DA (Data Access) is the original and most widely used standard of OPC. Its purpose is the cyclic polling of real time data, for example for visualization purposes.

• HDA (Historical Data Access), in contrary, specifies the access to already stored data. 

• AE (Alarms and Events) describes the non-cyclic, event-based exchange of alarms and events. 

• Data eXchange is a specification from 2002 which regulates the direct communication between two OPC servers. For this Master’s Thesis it was made use both of the DA specification for the main purpose of communication as well as the AE specification in order to display and log round-trip times. Unfortunately, the promising Data eXchange specification is almost inexistent in practice and could therefore not be used in our thesis. The underlying technique to exchange data is the component object model COM of Microsoft Windows, therefore OPC can only run on Windows operating systems [4]. A new generation of OPC specifications recently published is called OPC Unified Architecture (OPC UA) and is independent of COM, thus being able to run on more operating systems as well as embedded devices [5]. 

2.2.1 OPC Data Access OPC DA is organized in the hierarchical structure server, group and item. Items correspond to variables and can be read and written. Furthermore, a quality and time stamp is provided with each of them. When reading items, the value usually comes from the OPC server’s cache, which is updated periodically with the values of the device (or bus, component). However, it is usually possible to force a read directly from the device. Clients organize their items in groups, which for example share the same access method and update rate. Each OPC server has an unique name, some vendors even offer the operation of multiple servers for the same device. OPC DA provides different methods to access items, first of all synchronous and asynchronous read and write operations. More important to us, there is also a subscription mechanism, which is commonly used by modern clients in order to reduce communication. That is, the client group subscribes to the server which then “pushes” values towards the client only if they changed respectively exceed a pre-defined dead-band. The client can force an update of all these values by issuing a refresh call, which corresponds to an asynchronous read for all items of a group [6]. 

2.3 Programmable Logic Controllers This section informs about the two controllers involved and about the controller that has to be replaced. Please notice that we use the term controller equivalent to programmable logic controller (PLC) throughout our Master’s Thesis.