@article{4525,
  author = {Leonidas Kosmidis, Tullio Vardanega, Jaume Abella, Eduardo Quinones, Francisco J. Cazorla},
  title = {Applying Measurement-Based Probabilistic Timing Analysis to Buffer Resources},
  journal = {Journal of Electronic Systems},
  year = {2025},
  volume = {15},
  number = {3},
  doi = {https://doi.org/10.6025/jes/2025/15/3/161-174},
  url = {https://www.dline.info/jes/fulltext/v15n3/jesv15n3_3.pdf},
  abstract = {The paper explores the application of Measurement-Based Probabilistic Timing Analysis (MBPTA) to buffer
resources in Critical Real-Time Embedded Systems (CRTES). It establishes that buffers, unlike jittery resources
like caches, do not inherently create timing jitter; instead, they propagate existing jitter from other
resources. Buffers manage contention by decoupling request sending and processing speeds, and their behavior
is shown to be deterministic once probabilistic events are accounted for. The study demonstrates that
MBPTA can analyze buffers effectively as long as dependencies remain consistent between analysis and
operation. Empirical verification confirms that execution times are independent and identically distributed,
validating the suitability of MBPTA. Additionally, a classification of hardware resources based on jitter
sources is provided, enhancing the understanding of MBPTA compliance in processor architectures.},
}