Internet of Things (IoT) communications platform Real-Time Innovations has this week launched its latest Connext DDS 5.1 software platform to enable development and integration of distributed Intelligent Systems. "We are a messaging and integration middleware, very much targeted at Intelligent Systems," David Barnett, VP of Products and Markets at RTI, told ProgrammableWeb. "Our focus is on being the messaging and connection infrastructure in embedded machines. DDS is currently used in over 700 commercial applications. Our core domain is the intelligent-systems side."
Specific traits of Intelligent Systems
"Typically, you don't have systems administrators in Intelligent Systems, and they can't go down. There can be no single point of failure or even a failover point," Barnett says, describing some of the difficulties in managing IoT infrastructure at the "distributed Intelligent Systems" scale. These systems manage the data architecture backbone for machine-to-machine communications and machine-to-human interactions for aircraft fleets, medical and surgery technologies in hospitals, and energy utilities.
Managing the flow of data and communicating messages in real time is the core business of platforms seeking to service this market. RTI has been leveraging their DDS Connext platform to create a competitive advantage around the messaging and integration features they can offer to manage IoT infrastructures. Although messaging protocols in IoT have been taking up MQ Telemetry Transport (MQTT) and other pub/sub protocols, RTI has designed its own routing service, enhanced in Connext DDS 5.1.
"From the perspective of an application developer, the routing service is transparent, so it doesn't have an API," explained Barnett. "It is somewhat akin to a hardware Internet Protocol (IP) router, such as one from Cisco. Of course, it is a pure software solution and operates at the application layer instead of the network layer. That is, it is cognizant of the data within a message, so it makes decisions about which data to forward and where, based on the content and not just an address.
"An application just uses standard DDS APIs to send and receive data. Applications aren't aware of whether data is staying on the same network or being sent across a different network. When data is sent across network boundaries (for example, from a LAN to a WAN), the routing service decides whether to forward data or not, based on the data that is subscribed on the remote network. What's new in 5.1 is that the routing service automatically determines what data is subscribed, so it doesn't require any configuration or a special API.
"The routing service does have APIs for administration and customization. However, these are optional and not typically used by application developers. They would generally be used by system integrators, system administrators, or both."
Even as other market entrants in the IoT space (such as ClearBlade's Mobile Backend-as-a-Service) are seeking to use MQTT protocols to manage real-time communications for the IoT, RTI believes that their routing service is more suited to Intelligent Systems design.
"There are a couple of differences to MQTT," Barnett said. "We don't have a broker, so our routing service just monitors all data and selectively forwards the data that is needed. And it can dynamically change the data subscription: It can route messages based on a category or threshold condition because it is content aware."
An orchestration-type feature
One of the major challenges for API developers working with technologies in the IoT is the multiplicity of devices, each with their own nuances and integration needs. This has led Daniel Jacobson from Netflix to believe the future of API design will be in the orchestration layer. This sentiment has also led to next-generation solutions such as the recent launch of Orchestrate's single API to manage multiple databases.
The release of Connext DDS 5.1 applies an internal orchestration-type approach to managing distributed Intelligent Systems and the IoT. With new interoperability features, RTI is seeking to find a balance between enabling the introduction of new capabilities, while at the same time working with monolithic, legacy systems.
"You can evolve your data types without breaking any backward compatibility," said Barnett. "We do automatic mediation built-in to provide compatibility between different systems. You can support just one version of your data and have backward compatibility: It avoids having to publish data multiple times, and you use predefined rules as to how your data is understood by different vendors."
The interoperability feature is seen as a boon for developers building applications that integrate with existing Intelligent Systems.
"In Intelligent Systems, the deployment environment is not necessarily known, so another set of the enhancements in 5.1 is that it allows a protocol to auto-adjust, dependent on the network capability. This prevents the application from having to worry about that and greatly simplifies application development. It decreases much of the workload for the application developers and systems integrators, thus reducing the life-cycle costs that are essential with large machinery like airplanes and medical equipment.
"The backward-compatibility feature (Extensible Types) preserves existing API integrations in two ways:
- For applications that receive data using an old version of an API, it automatically strips out newer extensions when passing data to the application.
- For applications that receive data using a new version of an API, it indicates to the application which data is valid and which is unavailable, based on the version of the API (that is, the data type) being used by the sending application."
Building IoT capabilities into existing systems and ensuring efficient messaging from machine-to-machine and machine-to-human are two of the accelerators necessary for the true promise of a connected and enabled world to manifest. With Connext DDS 5.1, RTI believes it has solved these issues for one of the most crucial environments in the IoT landscape: Intelligent Systems.
By Mark Boyd. Mark is a freelance writer focusing on how we use technology to connect and interact. He writes regularly about API business models, open data, smart cities, Quantified Self, and e-commerce. He can be contacted via e-mail, on Twitter, or on Google+.