The Industrial Internet of Things: Integrating Industrial Systems with the Cloud
Clouds are elastic computing environments that can dynamically scale to meet a set of changing demands. These environments are becoming increasingly outsourced (public); they are viewed as the modern manifestation of IT, whether that is SaaS, PaaS or IaaS. The value of connecting operational industrial systems with cloud-based analytics and services is a major driver of the Industrial Internet of Things (IIoT).
The Cloud for the Consumer IoT is Different than the Cloud for the IIoT
There are many types of IoT systems. Most consumer IoT systems are simple; they connect a single device or a home gateway to a cloud data service. Thus, there is a fairly direct connection between a device and the cloud. But this architecture is not representative of most industrial systems. For instance, a manufacturing system may comprise hundreds of thousands of local sensors and processing nodes. While it may be possible to send all this information to the cloud, local intelligence that can triage or summarize the information is a compelling part of the system.
Cloud systems have three types of flows. Between devices and applications deployed at the "edge," between the edge and cloud (traditional M2M), and between applications in the cloud. RTI Connext DDS is uniquely capable of addressing all three requirements. This is a much greater challenge for the IIoT than for the consumer IoT.
Handling and Analyzing all Data at the Edge and in the Cloud
Today, most IIoT systems do not envision control in the cloud; they are restricted to monitoring and analysis. However, that is a limited view of the potential. Future IIoT systems will combine control, monitoring, and analysis.
Consider a smart city system that enables smart parking. It collects information across the city and provides it to the cloud, which makes that data available to smart phones and navigation systems. The cloud, and the mobile systems for that matter, are part of the system.
Another example is an ultrasound imaging system that combines real-time sensors, cloud-based image analytics, and connection with back-office IT. In order to be useful, the entire information trip from real-time sensor to cloud analytics and to the end-user (doctor) must be no more than a few milliseconds. As systems become faster, the entire network will be required to support real-time data flow.
Leveraging Public Clouds
Traditional industrial control system like robots, trains, or planes require data interactions on the order of milliseconds. It is not easy for public clouds to participate directly in this type of transaction because the network connections are too slow. However, private clouds can and should handle industrial grade transactions. RTI is currently researching radar track management in the cloud, one of the most demanding of RTI's applications. In the future, these systems may be able to leverage public cloud infrastructure.
Modular Design and Multi-Tiered Architecture
As industrial systems become increasingly complex, it is vital to build modern flexible architectures that are easy to update. RTI provides a bridging capability called RTI Routing Service that enables system integrators to compose complex systems from subsystems. RTI Routing Service provides two key functions: it allows subsystems to restrict which data is "exported," or visible, to higher-level systems. It also includes the ability to transform the data into the likely different structures used by higher-level systems. RTI Routing Service thus provides a way to build incrementally from parts while ensuring that the right data points will be accessible to the cloud.
IIoT Requires Integrated Security
Given the critical nature of most IIoT systems, security has to be a foremost concern. Traditionally, this has been a challenge because Operational Technology (OT) systems at the edge and IT system in the cloud used different connectivity technologies. Because RTI Connext DDS provides an edge-through-cloud platform, it also provides an end-to-end security solution encompassing authentication, access control and privacy.
Connext DDS Secure is also unique in that it directly supports fine-grained access control for data in motion without requiring a centralized message broker or Enterprise Service Bus (ESB). Other approaches are either limited to encrypting entire data channels or too complicated and require external access policies and roles that map poorly to IIoT systems. RTI Connext simply enforces control of the actual data flow as specified by system designers. This design prevents intentional sabotage and hacking as well as unintended human error that could also compromise industrial system and cause safety issues. It is extremely fast and scalable, even including support for multicast security.
DDS security model simplifies security of IIot and makes it easy to manage, without compromising elegance and power. For example, a hospital system may need to ensure that only an authorized care team may access a particular patient's data. With coarse-grained security, this is a very large project. With DDS Secure, it simply requires a specification of the dataflow relationships—what application should get what data. It requires no specialized security expertise and no coding skills.
Eze Software Group
RTI Queuing Service and Connext DDS technology are at the core of a new cloud-enabling platform that facilitates highly scalable, resilient and secure data exchange between multiple Eze Software products and services. Eze selected RTI for its industry-leading, fully DDS-compliant publish-subscribe implementation.
Eze is using DDS to collect information from possibly thousands of traders, and to process the trades and report the information back to system users. The Queueing Service allows cloud-based analytics to process vast amounts of information by distributing loads across many elastic systems. This will also run on private clouds, and even incorporate load-balancing hardware such as the F5 system. This a very typical architecture for large websites that handle many simultaneous interactions. The difference with DDS is the ability to coordinate the responses and respond quickly to the data sources.