Conference Papers

Permanent URI for this collectionhttps://idr.nitk.ac.in/handle/123456789/28506

Browse

Filters

2008 - 200912020 - 20241

Settings

Subject: search.filters.subject.Deadline

Search Results

Now showing 1 - 2 of 2
  • No Thumbnail Available
    Item
    Development of scheduler for real time and embedded system domain
    (2008) Rao, M.V.P.; Shet, K.C.; Balakrishna, R.; Roopa, K.
    We discuss scheduling techniques to be used for real-time, embedded systems. Though there are several scheduling policies, the preemptive scheduling policy holds promising results. In this research paper, the different approaches to design of a scheduler for real-time Linux kernel are discussed in detail. The comparison of different preemptive scheduling algorithms is performed. Hence, by extracting the positive characteristics of each of these preemptive scheduling policies, a new hierarchical scheduling policy is developed. The proposed hierarchical scheduling for real time and embedded system will be implemented for a prototype system, using C or C++ language. It is expected that the new scheduling algorithm will give better performance with respect to satisfy the needs, such as time, capturing and usage of resources of different applications. © 2008 IEEE.
  • No Thumbnail Available
    Item
    LEASE: Leveraging Energy-Awareness in Serverless Edge for Latency-Sensitive IoT Services
    (Institute of Electrical and Electronics Engineers Inc., 2024) Verma, A.; Satpathy, A.; Das, S.K.; Addya, S.K.
    Resource scheduling catering to real-time IoT services in a serverless-enabled edge network is particularly challenging owing to the workload variability, strict constraints on tolerable latency, and unpredictability in the energy sources powering the edge devices. This paper proposes a framework LEASE that dynamically schedules resources in serverless functions catering to different microservices and adhering to their deadline constraint. To assist the scheduler in making effective scheduling decisions, we introduce a priority-based approach that offloads functions from over-provisioned edge nodes to under-provisioned peer nodes, considering the expended energy in the process without compromising the completion time of microservices. For real-world implementations, we consider a testbed comprising a Raspberry Pi cluster serving as edge nodes, equipped with container orchestrator tools such as Kubernetes and powered by OpenFaaS, an open-source serverless platform. Experimental results demonstrate that compared to the benchmarking algorithm, LEASE achieves a 23.34% reduction in the overall completion time, with 97.64% of microservices meeting their deadline. LEASE also attains a 30.10% reduction in failure rates. © 2024 IEEE.