Ensuring Reliable System Operation in Challenging Network Environments

From vehicles and medical devices to industrial equipment and defense platforms, the shift to software-defined systems is already well underway.  

The obstacle? Reliable data exchange across unreliable networks. In the field, whether navigating congested city streets, coordinating medical devices in a hospital, or operating in contested environments, networks are rarely perfect. Bandwidth is limited, interference is common, and connections drop without warning. Without a way to ensure data reaches every participant, collaboration between systems breaks down, creating risks for safety, performance, and compliance.

This is why reliability matters: 

• Reliable data exchange ensures that systems can act in unison rather than as isolated devices
• Decisions can be made with confidence based on complete and consistent information, and 
• Operations can continue seamlessly even when connectivity falters. 

The solution begins with a data-centric architecture powered by RTI Connext, strengthened by the new synchronization features in RTI Persistence Service to ensure that distributed systems remain reliable, consistent, and ready to operate as one.

Persistence Service as the Foundation of Reliability

At the heart of every distributed system lies a simple need: data must be available when and where it is needed. RTI Persistence Service addresses this with three core capabilities:

1. Relays data from lost applications: Persistence Service manages relevant Data Distribution Service (DDS^™) data samples, making them available to late-joining subscribing applications even if the original publishing application is no longer running. 
2. Keeps the system consistent: Persistence Service can run across multiple instances that share the same system view. If one misses a piece of data, it can quickly pull it from another.
3. Preserves data for post-mortem analysis: The service also offers configurable synchronization levels with the physical disk, such as a persistent mode, where data survives service restarts, further enhancing fault tolerance for critical systems. 

What’s New: Synchronization Across Instances

The latest enhancement adds synchronization across Persistence Service instances, a major step forward in resilience. This feature uses a gossip-style approach to reconcile data across Persistence Service instances, ensuring that over time, all nodes converge on the same consistent set of information. The result is resilient and efficient synchronization of stored data that doesn’t depend on a central broker or constant cloud connectivity.

The compelling advantage of Persistence Service's new synchronization mechanism is that each instance shares a summary of the data it possesses along with the actual data. By checking this information before sending the full data, redundant transmissions are avoided, ensuring efficient and optimized data propagation across the network. 

Example scenario: 

Consider a network of devices operating in a harsh or bandwidth-constrained environment, e.g. smart tractors in a field, drones in contested airspace, or autonomous vehicles navigating city streets. In each case, a temporary disconnection can drop critical data, putting safety, performance, or mission outcomes at risk.

This is where the synchronization capability in Persistence Service can help applications achieve eventual consistency with their data. Each Persistence Service instance exchanges updates with its discoverable Persistence Service neighbors, which then propagate missing data further, until all Persistence Service instances in the system converge on the same data. This enables any device – whether a tractor, drone, or vehicle – to eventually obtain data generated by another, even if that original source is temporarily unreachable or no longer active. All of this happens without requiring a centralized server or permanent connectivity.

Figure 1, below, reflects the visual comparison between two options. Without Persistence Service, data is only delivered to devices within direct range, leaving others unaware of critical threats. With synchronized Persistence Service, data is propagated across devices over time, ensuring eventual consistency and shared situational awareness, even in unreliable networks.

Figure 1.

Enabling future operations

The synchronization capabilities of RTI Persistence Service provide a powerful way to keep distributed systems reliable, even when networks are unreliable. By sharing data efficiently among peers, systems achieve eventual consistency, ensuring that no critical information is lost and that operations can continue without interruption. 

Combined with RTI Connext’s strengths in secure data handling, real-time performance, and scalability, Persistence Service gives organizations the confidence to deploy distributed systems in the most challenging environments.

From defense to healthcare to industrial markets, this enhancement empowers customers to accelerate development, reduce risk, and future-proof their operations. It’s not just about moving data  it’s about ensuring reliability and resilience where it matters most.

Explore our Github example on Persistence Service and see it in action. New to Connext? Download the free trial and begin building more resilient distributed applications.

About the author:

Andrés Romero is an Application Engineer within the RTI Professional Services team. He works closely with a wide variety of RTI's customers to help them design their systems using Connext. His experience includes customer use cases across defense, healthcare, and industrial IoT sectors.