Document Type

Article

Source of Publication

IEEE Access

Publication Date

1-1-2024

Abstract

Real-time systems mostly interact with the external world and each input operation must meet predetermined deadlines to be useful. However, in many real-time applications, a partial result is also acceptable. We developed a reward-based mixed criticality system based on the resource reservation approach to address the problem of ensuring the effective execution of low- and high-criticality tasks in both low- and high modes, even under heavy workloads. Using dedicated servers with pessimistic resource allocation for each high criticality task ensured their execution in both modes unaffected by low criticality tasks. The surplus resources are reclaimed and assigned to low critical tasks' server by utilizing a greedy reclamation of unused bandwidth (GRUB) algorithm. Three strategies were suggested for server allocation to low criticality tasks: a dedicated server for all low criticality tasks, a single server for each low criticality task, and two servers (mandatory and optional) for each low criticality task. Our analysis revealed efficiency of the first approach by achieving 100% schedulability at a 1.1 target utilization, scheduling 20% and 50% more task sets than the second and third approaches, respectively. Moreover, the effectiveness of the proposed approach over existing imprecise mixed criticality approaches were demonstrated through comprehensive experimentation.

ISSN

2169-3536

Publisher

Institute of Electrical and Electronics Engineers (IEEE)

Volume

12

First Page

49134

Last Page

49150

Disciplines

Computer Sciences

Keywords

Greedy reclamation, imprecise computation, imprecise mixed criticality (IMC) systems, mixed criticality systems, real-time systems, resource reservation reward-based scheduling, servers

Scopus ID

85189645652

Indexed in Scopus

yes

Open Access

yes

Open Access Type

Gold: This publication is openly available in an open access journal/series

Download

Share

COinS