SSSepideh Safari

Sepideh Safari

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Sepideh Safari has a total of 33 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • TORE-NET: A Task Replication and Offloading Strategy in Reliable and Energy-Efficient Mobile Edge Computing Networks

    by Amir Mahdi Rasouli, Moein Esnaashari, Abbas Shabrang Maryan, Mohsen Ansari, Sepideh Safari, Muhammad Shafique

  • Pilot: P ower-Aware Hybr i d Fault To l erance in Multi-C o re Embedded Sys t ems

    by Amir Hossein Ansari, Moein Esnaashari, Sepideh Safari, Mohsen Ansari, Alireza Ejlali, Jörg Henkel

    With the advancement of technology size and the integration of multiple cores on a single chip, the probability of fault occurrence has increased. These faults can be transient or permanent, requiring techniques to manage both types. Hybrid fault tolerance techniques have emerged as effective solutions to handle both types. In this paper, we propose a power-aware hybrid fault tolerance (called Pilot). Our approach utilizes checkpointing with rollback-recovery and primary/backup techniques, tolerating two kinds of faults. Moreover, in real-time embedded systems, power consumption is a critical constraint that must be managed. To do this, we exploit the Thermal Safe Power (TSP) constraint for each processing core. Based on this constraint and the utilization of each core, tasks are mapped and scheduled, while guaranteeing the timing constraints. Our experimental results demonstrate that our proposed methods can meet the reliability target by tolerating the optimal number of fault occurrences in each task while reducing power consumption. Our proposed methods are compared to state-of-the-art techniques in terms of schedulability, power consumption, Quality of Service (QoS), energy consumption, and reliability. The peak power and energy consumption are reduced on average by 34.2% and 15.9%, respectively, the QoS is improved on average to 28.7%, and the schedulability is improved on average to 14.6% while satisfying the system reliability target.


  • RTGPU: Real-Time Computing with Graphics Processing Units

    by Atiyeh Gheibi-Fetrat, Sepideh Safari, Shaahin Hessabi, Hamid Sarbazi-Azad

  • A Two-State Energy-Efficient Reliability-Aware Strategy in Embedded Systems

    by Amir Mahdi Rasouli, Amir Hossein Ansari, Moein Esnaashari, Sepideh Safari, Mohsen Ansari

    Reducing energy consumption while tolerating transient faults is a significant challenge in modern embedded systems with weakly-hard real-time requirements. In this paper, we present NB-GEAR, a two-state runtime algorithm that combines Dynamic Voltage and Frequency Scaling (DVFS) with non-uniform checkpointing to save energy in normal operation and switches to a high-reliability mode only when the system can no longer tolerate further deadline misses. NB-GEAR preserves both the schedulability and reliability target of the system while achieving up to 15.1% energy savings compared to other methods. Experiments on MiBench benchmarks in a gem5+McPAT setup confirm that our method delivers real-time guarantees, reliability, and energy savings in one cohesive solution.

  • Open Access

    MOSAIC: Mobility-Oriented Scheduling and Intelligent Resource Allocation for IoT

    by Abolfazl Younesi, Mehrab Toghani, Sepideh Safari, Mohsen Ansari, Thomas Fahringer

    The relentless growth of mobile Internet of Things (IoT) devices has shifted computation toward a distributed computing continuum, spanning edge, fog, and cloud layers, where energy efficiency, low latency, and dynamic node mobility are critical yet often conflicting goals. Existing scheduling frameworks struggle to balance these demands under real-world conditions, especially as device movement and heterogeneous workloads increase system complexity. We present MOSAIC, a mobility-aware scheduling and resource management framework designed to optimize performance in dynamic IoT environments. Our approach introduces three key innovations. First, a refined five-tier architecture extends the traditional edge-fog-cloud hierarchy by adding proximity, local, and regional mobility layers, enabling computation to follow mobile users more effectively and reducing unnecessary network traffic. Second, MOSAIC integrates a preemption-aware dynamic scheduler with an Adaptive-$\lambda$ reinforcement learning-based resource manager that adapts based on workload changes and mobility patterns, prioritizing energy-efficient edge execution while meeting strict deadlines. Third, the framework utilizes real-world mobility traces, including Levy-Walk, Random-Walk, and Geolife, to drive reconfiguration and improve decision accuracy. We evaluate MOSAIC through a large-scale deployment across three geographically distributed regions of the Grid’5000 testbed, using realistic workflows and mixed periodic/DAG task loads. Our results show that, compared to state-of-the-art schedulers, MOSAIC reduces energy consumption by 35.9% –×1.5, lowers latency by 42.8% –×4.9, and shortens makespan by 22.6% –×7.2, all while maintaining 100% deadline satisfaction across diverse mobility scenarios.


  • Padel: Priority-Based Real-Time Scheduling for GPUs

    by Atiyeh Gheibi-Fetrat, Sepideh Safari, Amirsaeed Ahmadi-Tonekaboni, Shaahin Hessabi, Hamid Sarbazi-Azad

    Graphics Processing Units (GPUs) have become increasingly prevalent in many platforms, including real-time systems, due to their massive architectural parallelism and significant performance. Power and energy management, reducing deadline miss rate (DMR), and efficient allocation of resources to tasks are important design challenges in exploiting GPUs within real-time platforms. One of the most considerable challenges in designing firm real-time systems is allocating GPU resources to tasks in a way that enables as many tasks as possible to be completed correctly within their deadlines while minimizing energy consumption. However, since multiple tasks can be handled concurrently across many cores, power consumption poses a serious limitation. In this work, we introduce Padel, a real-time GPU scheduler that employs spatial multitasking to enhance the utilization of resources, performance, and energy efficiency in GPUs. Our experimental results reveal that when the system is overloaded, Padel can reduce the DMR by 23%, and energy consumption by 8% compared to the state of the art.


  • DIST: Distributed Learning-Based Energy-Efficient and Reliable Task Scheduling and Resource Allocation in Fog Computing

    by Elyas Oustad, Abolfazl Younesi, Mohsen Ansari, Sepideh Safari, Mohammad Arman Soleimani, Jörg Henkel, Alireza Ejlali

    This paper presents DIST, a novel distributed reinforcement learning-based (DRL) framework for energyefficient and reliable task scheduling and resource allocation in fog computing, low-latency computing solutions driven by the rapid deployment of IoT devices, and time-sensitive applications. DIST is built based on a novel distributed Q-learning to enable fog nodes to learn an optimal strategy to balance energy consumption, task execution time, and system reliability. The main novelty includes a cooperative Dynamic Voltage and Frequency Scaling-enabled task scheduling policy that dynamically adjusts node energy level to ensure power consumption reduction without sacrificing deadline adherence or reliability. The results demonstrate that DIST reduces energy consumption by up to 52.26%, realizes 38% higher success rates, and reduces task wait times by up to 46.77%, compared with state-of-the-art algorithms.


  • Open Access

    ReTMiC: Reliability-Aware Thermal Management in Multicore Mixed-Criticality Embedded Systems

    by Sepideh Safari, Mohsen Ansari, Shaahin Hessabi, Jörg Henkel

    As the number of cores in multicore platforms increases, temperature constraints may prevent powering all cores simultaneously at maximum voltage and frequency level. Thermal hot spots and unbalanced temperatures between the processing cores may degrade the reliability. This paper introduces a reliability-aware thermal management scheduling (ReTMiC) method for mixed-criticality embedded systems. In this regard, ReTMiC meets Thermal Design Power as the chip-level power constraint at design time. In order to balance the temperature of the processing cores, our proposed method determines balancing points on each frame of the scheduling, and at run time, our proposed lightweight online re-mapping technique is activated at each determined balancing point for balancing the temperature of the processing cores. The online mechanism exploits the proposed temperature-aware factor to reduce the system’s temperature based on the current temperature of processing cores and the behavior of their corresponding running tasks. Our experimental results show that the ReTMiC method achieves up to 12.8°C reduction in the chip temperature and 3.5°C reduction in spatial thermal variation in comparison to the state-of-the-art techniques while keeping the system reliability at a required level.


  • Open Access

    LEC-MiCs: Low-Energy Checkpointing in Mixed-Criticality Multicore Systems

    by Sepideh Safari, Shayan Shokri, Shaahin Hessabi, Pejman Lotfi-Kamran

    With the advent of multicore platforms in designing Mixed-Criticality Systems (MCSs), simultaneous management of reliability and energy while guaranteeing an acceptable service level for low-criticality tasks is a crucial challenge. To ensure the reliability of the MCSs against transient faults, fault-tolerant techniques are employed which will increase energy consumption. To mitigate the energy overhead, the Dynamic Voltage and Frequency Scaling (DVFS) technique will be exploited. However, this technique might lead to violating the timing constraints of high-criticality tasks. Therefore, this article presents, for the first time, the low-energy checkpointing technique to guarantee the reliability of multiple preemptive periodic mixed-criticality tasks in a multicore platform. In contrast to the previous works in checkpointing technique which consider a specific number of faults that all the tasks in the system should tolerate, in this article, the number of tolerable faults for each execution section of a task and in each voltage and frequency level is determined through proposed formulas to meet the reliability target based on safety standards. Then, our proposed method determines the number of checkpoints and their non-uniform intervals for the normal and overrun sections of each task to reduce energy consumption, respectively. Moreover, the unified demand bound function (DBF) analysis is proposed for analyzing the schedulability of the task set, where each high-criticality task meets its timing and reliability constraints, and low-criticality tasks execute based on their derived guaranteed periods in each operational mode of the system. Experimental results show that our proposed scheme meets the timing and reliability constraints while at the same time, improving the Quality of Service (QoS) of low-criticality tasks and managing energy consumption with an average of 29.49% and 32.78%, respectively.


  • Energy-Aware Fault-Tolerant Mapping of Mixed-Criticality Tasks on Heterogeneous Multicores

    by Amir Hassan Safizadeh, Sepideh Safari, Shayan Shokri, Shaahin Hessabi

    Due to the different timing requirements of tasks in different criticality modes, it is challenging for Mixed-Criticality Systems (MCSs) designers to use time-redundant fault-tolerant techniques. Checkpointing with rollback recovery can be an effective option for ensuring the reliability of tasks in such systems. Despite the benefits of checkpointing it can impose significant energy overheads. A common approach to overcome this energy consumption is using Dynamic Voltage and Frequency Scaling (DVFS). However, DVFS can lead to missed deadlines for high-criticality tasks. In addition, there is an increasing trend of deploying tasks on heterogeneous multicore platforms. These platforms offer promising opportunities to meet the design requirements of mixed-criticality applications more effectively. In this paper, we propose an energy-aware checkpointing scheme for mixed-criticality tasks on heterogeneous multicore platforms. First, we calculate the timing demand of each task according to DVFS and the time overhead of checkpointing to perform the EY schedulability test in different operational modes of the system. Then we focus on energy-aware mapping of tasks to heterogeneous platforms. Finally, we use DVFS to mitigate the energy overhead of checkpoints. Our experiments show that our scheme can improve schedulability by an average of 16% compared to Little Island First (LIF) mapping, while the energy does not change appreciably. Moreover, it consumes up to 36% (20% on average) less energy compared to Big Island First (BLF) mapping.