2015/01/01 by Enrico Mezzetti, Mezzetti, Enrico, Marco Ziccardi +9
Computer Science · #Computer systems organization → Real-time system architecture #Computer systems organization → Special purpose systems #Distributed systems and fault tolerance #Parallel Computing and Optimization Techniques #Probabilistic WCET #Randomized caches #Real-Time Systems Scheduling #Real-time systems #Theory of computation → Probabilistic computation
paper · doi:10.4230/lites-v002-i001-a001
openalex publication_date 2015/01/01 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Cache randomization per se, and its viability for probabilistic timing analysis (PTA) of critical real-time systems, are receiving increasingly close attention from the scientific community and the industrial practitioners. In fact, the very notion of introducing randomness and probabilities in time-critical systems has caused strenuous debates owing to the apparent clash that this idea has with the strictly deterministic view traditionally held for those systems. A paper recently appeared in LITES (Reineke, J. (2014). Randomized Caches Considered Harmful in Hard Real-Time Systems. LITES, 1(1), 03:1-03:13.) provides a critical analysis of the weaknesses and risks entailed in using randomized caches in hard real-time systems. In order to provide the interested reader with a fuller, balanced appreciation of the subject matter, a critical analysis of the benefits brought about by that innovation should be provided also. This short paper addresses that need by revisiting the array of issues addressed in the cited work, in the light of the latest advances to the relevant state of the art. Accordingly, we show that the potential benefits of randomized caches do offset their limitations, causing them to be - when used in conjunction with PTA - a serious competitor to conventional designs.