Unreliable failure detectors for reliable distributed systems
Journal of the ACM (JACM)
The Timed Asynchronous Distributed System Model
IEEE Transactions on Parallel and Distributed Systems
Simulation Modeling and Analysis
Simulation Modeling and Analysis
Distributed Systems for System Architects
Distributed Systems for System Architects
Measuring Distributed Durations with Stable Errors
RTSS '01 Proceedings of the 22nd IEEE Real-Time Systems Symposium
A Freshness Detection Mechanism for Railway Applications
PRDC '04 Proceedings of the 10th IEEE Pacific Rim International Symposium on Dependable Computing (PRDC'04)
International Journal of Critical Computer-Based Systems
| Hi-index | 0.00 |
The detection of messages delayed more than a given threshold – called 'message freshness detection' – is an important requirement in many distributed critical real-time systems. In this paper, a solution for addressing this requirement is described and deeply analysed. Our solution, the 'available and safe freshness detection algorithm (ASFDA)', was proposed for a distributed infrastructure for railway control, but it fits also to different systems with similar characteristics. ASFDA allows the detection of 'old' messages: violations of real-time requirements. ASFDA is based on round trip time estimation technique, a well known method used in distributed systems with real-time requirements. In the paper, we demonstrate that ASFDA can detect all the messages older than allowed; we also demonstrate, through an availability analysis performed by simulation, that the availability penalty reached by the proposed freshness detection algorithm is practically negligible if compared with its theoretical limits.