Azure Gateway DataRepository
Azure, Cloud Computing, Object-Oriented Internet, OPC Unified Architecture, Reactive Networking (RxNetworking), Machine to Machine Communication, Internet of Things
This project shows an example implementation of an OPC UA PubSub to Azure embedded gateway. It is implemented as a composable part of the Reactive Networking Application (RxNetworking App). The article Reactive Networking of Semantic-Data Library covers a description of the architecture supporting the reactive communication design pattern. The RxNetworking App is an aggregation of Producer and Consumer entities derived from DataRepository. They must provide interconnection to real-time process data, hence they are recognized as an extension of the DataRepository class. AzureGateway part fulfills the Consumer role and uses out-of-band communication to push telemetry data to the cloud.
Working through this tutorial gives you an introductory understanding of the steps required to implement the Consumer role of the RxNetworking application. It is an example of Semantic-Data reactive networking based on OPC UA PubSub specification. The document OPC UA PubSub Main Technology Features covers a description of selected fetuses relevant to this specification.
This project is proof of concept that out-of-band communication for OPC UA PubSub can be implemented based on the DataRepository concept. This workout will be described in an independent article. To get the full story and your copy check out the preprint from Research Gateway: Object-Oriented Internet - Azure Interoperability. Main purpose of this preprint it to enable an early community review. We will consider your contribution to be applied to the final version of the article.
The obtained results prove that the embedded gateway archetype implementation is possible based on the existing standalone framework supporting reactive interoperability atop of the M2M communication compliant with the OPC UA PubSub standard. It is worth stressing that there is no dependency on the Client/Server session-oriented relationship. In contrast to the architecture described in the OPC UA Part 1 specification, the publisher/consumer roles are not tightly coupled with the Address Space of the OPC UA Server embedded component. In the proposed approach, the cloud interoperability is supported by a dedicated part employing out-of-band communication only without dependency on the OPC UA functionality. In contrast to the middleware concept, the gateway functionality is implemented as a part - composable to the whole without programming skills. It makes it possible to modify its functionality later after releasing the library or even deploying the application program in the production environment because the part is composed at the runtime.
Concluding, the paper describes a proof of concept that it is possible to integrate selected cloud services (e.g. Azure) with the Cyber-physical network atop of the OPC UA PubSub applying the proposed architecture and deployment scenario. In contrast to limiting the PubSub role to export the data from the Address Space exposed by a selected OPC UA server out of the OPC UA ecosystem, applying the proposed solution enables interoperability of the cloud services and the Cyber-physical network as one whole.
I would like to acknowledge the CrossHMI project from which the AzureGateway implementation of the DataRepository was derived. I would like to thank Piotr Szymczak for his inputs/inspirations, feedback, and cooperation in this respect.
Here are steps undertook to implement the Consumer role in the application:
DataManagementSetup: this class has been overridden by thePartDataManagementSetupclass and it initializes the communication and binds data fields recovered form messages to local resources.IEncodingFactoryandIMessageHandlerFactory: have been implemented in external common libraries andConsumerdoesn't depend on this implementation - current implementation of the interfaces is localized as services using an instance of the CommonServiceLocator.IServiceLocator interface.IBindingFactory: has been implemented in the classPartBindingFactorythat is responsible to gather the data recovered from theMessageinstances pulled from theDistribution Channel. The received data is driven to the Azure services using configured out-of-band protocol.IConfigurationFactory: the classPartConfigurationFactoryimplements this interface to be used for the configuration file opening.
The PartDataManagementSetup constructor initializes all properties, which are injection points of all parts composing this role.
[Export(typeof(PartDataManagementSetup))][PartCreationPolicy(CreationPolicy.Shared)]public sealed class PartDataManagementSetup : DataManagementSetup{public PartDataManagementSetup(){_Logger.EnteringMethodPart(nameof(PartDataManagementSetup));//Compose external partsIServiceLocator _serviceLocator = ServiceLocator.Current;//string _configurationFileName = _serviceLocator.GetInstance<string>(CompositionSettings.ConfigurationFileNameContract);m_ViewModel = _serviceLocator.GetInstance<ProducerViewModel>();EncodingFactory = _serviceLocator.GetInstance<IEncodingFactory>();_Logger.Composed(nameof(EncodingFactory), EncodingFactory.GetType().FullName);MessageHandlerFactory = _serviceLocator.GetInstance<IMessageHandlerFactory>();_Logger.Composed(nameof(MessageHandlerFactory), MessageHandlerFactory.GetType().FullName);//compose internal partsConfigurationFactory = new PartConfigurationFactory(ConfigurationFilePath);PartBindingFactory pbf = new PartBindingFactory();_DTOProvider = pbf;BindingFactory = pbf;}....}
In this example, it is assumed that IServiceLocator is implemented to resolve references to any external services.
Finally the DataManagementSetup.Start() method is called to initialize the infrastructure, enable all associations and start pumping the data.
Implementation of this interface is a basic step to implement Consumer functionality. The DataRepository represents data holding assets in the RxNetworking App and, following the proposed approach, the IBindingFactory interface is implemented by an external part. It captures functionality responsible for accessing the process data represented by the LocalResources. The LocalResources represents the external part that has a very broad usage purpose. For example, it may be any kind of process data source/destination, and to name a few Raw Data, OPC UA Address Space Management, Azure cloud-based front-end, etc.
The AzureGateway functional package has been implemented based on the Consumer concept. This particular Consumer (PartBindingFactory) implements the IBindingFactory interface to gather the data recovered from the Message instances pulled from the Distribution Channel. The received data is driven to the Azure services using configured out-of-band' protocol. An instance of the IBindingFactory is responsible to create objects implementing IBinding that can be used by the Consumer to forward the data retrieved from NetworkMessag received over the wire to Azure services.
The proposed implementation of the Azure gateway proves that the DataRepository and associated entities, i.e. Local Resources, Consumer, Producer can be implemented as external parts, and consequently, the application scope may cover practically any concern that can be separated from the core OPC UA PubSub communication engine implementation.
the library UAOOI.Configuration.Networking. In a typical scenario, this implementation should not be considered for further modification. The only open question is how to provide the name of the file containing the configuration of this role. This role uses an independent configuration file:
ConfigurationDataConsumer.BoilersSet.xml
attached to the project.
Note; This library is not considered to be published as the NuGet package.
We use Semantic Versioning 2.0.0 for versioning. For the versions available, see the Releases page of the project.
Mariusz Postol - main contributor of this project.
See also the list of contributors who participated in this project and the Acknowledgment section.
Postół M. (2020) Object-Oriented Internet Reactive Interoperability. In: Krzhizhanovskaya V. et al. (eds) Computational Science – ICCS 2020. ICCS 2020. Lecture Notes in Computer Science, vol 12141. Springer, Cham; DOI: https://doi.org/10.1007/978-3-030-50426-7_31
Postół M. (2020) Object-Oriented Internet Reactive Interoperability, presentation, DOI: 10.13140/RG.2.2.33984.56323
Mariusz Postol, Machine to Machine Semantic-Data Based Communication: Comprehensive Survey chapter in book Computer Game Innovations 2018, Publisher: Lodz University of Technology Press; ISBN: 978-83-7283-999-2
API Browser: the preliminary code help documentation - available for sponsors- consider joining
