TTTECH Labs has received another major recognition for its research in deterministic real-time systems. At ECRTS 2026, one of the leading international conferences in the field, the paper "Multi-Core Integration of Sporadic Events in Time-Triggered Systems" was selected as an Outstanding Paper.
The award recognizes research that addresses a growing challenge in modern computing platforms: how to make increasingly complex multicore systems more efficient while maintaining strict real-time guarantees.
Before exploring the research itself, it is worth considering a simple thought experiment.
Imagine a busy restaurant on a Saturday evening. Some activities are carefully planned in advance, while others emerge unexpectedly throughout the night. The key to a successful operation is not just an efficient plan - it is leaving enough flexibility to respond to the unexpected without affecting overall performance.
Keeping this image in mind provides an intuitive way to think about the challenge explored by Anaïs Finzi and Silviu S. Craciunas. Although the systems they study are far more complex than a restaurant, they face a remarkably similar question: how can resources be organized so that planned activities and unforeseen events can coexist efficiently without compromising reliability?
TTTECH Labs continues to advance the state of the art in real-time systems research. At the 38th European Conference on Real-Time Systems (ECRTS 2026) in Lund, Sweden, the paper “Multi-Core Integration of Sporadic Events in Time-Triggered Systems” was selected as one of four Outstanding Papers at the conference. The paper was co-authored by Anaïs Finzi, Senior Solution Architect at TTTECH Labs, and Silviu S. Craciunas, who is currently affiliated with NXP Semiconductors and the Technical University of Denmark (DTU).
The recognition marks another important achievement for TTTECH Labs and builds on the team's previous award-winning research presented at RTAS 2024, where a novel approach for integrating time-triggered and sporadic event-triggered workloads in real-time systems received an Outstanding Paper Award.
As computing platforms in domains such as automotive, aerospace, and cloud infrastructure continue to grow in complexity, increasing performance is no longer simply a matter of adding more hardware. More processing cores, larger networks, and additional system resources also increase cost, power consumption, weight, and system complexity. As a result, improving the efficiency of existing computing platforms has become a key research challenge.
The ECRTS 2026 paper addresses one of the fundamental challenges in modern real-time systems: enabling time-triggered and event-triggered workloads to coexist efficiently on multicore platforms. While time-triggered tasks follow a predetermined schedule, event-triggered tasks occur dynamically in response to external events. Scheduling decisions made during system design can therefore have a direct impact on whether event-triggered workloads will later be able to meet their deadlines.
This insight is at the heart of the paper’s contribution. Rather than creating a time-triggered schedule in isolation and analyzing the impact on event-triggered tasks afterwards, the proposed approach incorporates the requirements of event-triggered workloads from the outset. In other words, the time-triggered schedule is generated in a way that deliberately preserves sufficient computing capacity for dynamically occurring tasks, ensuring that both execution paradigms can coexist while maintaining their respective timing guarantees.
“Traditionally, time-triggered schedules and event-triggered workloads are often considered separately. Our approach brings them together from the very beginning of the design process. By accounting for the timing requirements of event-triggered tasks during schedule generation, we can utilize multicore resources more efficiently while still guaranteeing predictable system behaviour. This becomes increasingly important as safety-critical platforms continue to grow in complexity and software content.”
Anais Finzi, Senior Solution Architect at TTTECH Labs
A key advancement over the earlier RTAS research is the target architecture. The previous work focused on fully partitioned multicore systems, where tasks are assigned to specific processor cores before scheduling and cannot migrate between them. The new research extends the concept to global multicore systems, where tasks can be assigned across processor cores during schedule generation instead of being permanently tied to a single core. This additional flexibility increases the complexity of scheduling but also creates new opportunities to utilize computing resources more efficiently while preserving deterministic real-time behaviour.
By successfully transferring the concept from fully partitioned to global multicore architectures, the authors demonstrate how deterministic real-time behaviour can be maintained while enabling a more flexible and scalable use of multicore processing resources. This evolution represents an important step toward supporting increasingly sophisticated real-time workloads on next-generation computing platforms.
To achieve this, the authors developed a method that combines formal timing analysis and advanced scheduling techniques to generate time-triggered schedules that guarantee the schedulability of event-triggered tasks. By extending existing timing-analysis models and adapting scheduling strategies for global multicore environments, the approach enables efficient resource sharing between both task types while maintaining strict timing guarantees.
Experimental results demonstrate the effectiveness of the approach. The proposed global scheduling methods outperform existing global approaches in terms of feasibility rates and average runtime. The evaluation also showed that Response Time Analysis (RTA) delivers the same feasibility results as Real-Time Calculus (RTC) while requiring significantly less computation time. In the evaluated real-world test cases, RTA completed in fractions of a second, whereas RTC required several hundred seconds.
The research further indicates that the proposed method scales well as the number of processor cores increases. For moderate system utilization levels, the global multicore approach can outperform existing fully partitioned methods and make more efficient use of available computing resources. According to the authors, this can enable the same workload to be executed using fewer CPUs compared to existing approaches, potentially reducing system weight, power consumption, complexity, and overall cost.
The findings are particularly relevant for future safety-critical computing platforms, where ever-increasing software complexity must be balanced with strict requirements for determinism, efficiency, and scalability. By allowing resources to be utilized more effectively while preserving timing guarantees, the proposed approach contributes to the development of next-generation real-time architectures for demanding applications.
“Future software-defined vehicles, aerospace platforms, and industrial systems will require significantly more computing power while maintaining the highest levels of safety and predictability. Research like this helps ensure that we can meet these demands without simply adding more hardware. By enabling a more efficient use of multicore resources, we can support increasingly sophisticated applications while keeping real-time guarantees intact.”
Anais Finzi, Senior Solution Architect at TTTECH Labs
By extending a previously award-winning concept from fully partitioned to global multicore architectures, Anaïs Finzi and Silviu S. Craciunas have demonstrated a promising path toward more efficient and scalable real-time computing platforms. The Outstanding Paper recognition at ECRTS 2026 highlights the continued contribution of TTTECH Labs to advancing deterministic computing technologies for next-generation safety-critical systems.
