DLJSF: Data-Locality Aware Job Scheduling IoT tasks in fog-cloud computing environments

• Published in: Results in engineering 2024-03, Vol.21, p.101780, Article 101780
• Main Authors: Khezri, Edris, Yahya, Rebaz Othman, Hassanzadeh, Hiwa, Mohaidat, Mohsen, Ahmadi, Sina, Trik, Mohammad
• Format: Article
• Language: English
• Subjects: Cloud computing; Fog computing; Internet of things; Locality-aware; Optimisation; Scheduling
• ISSN: 2590-1230; 2590-1230
• DOI: 10.1016/j.rineng.2024.101780

Nowadays, devices generate copious quantities of high-speed data streams due to Internet of Things (IoT) applications. For the most part, cloud computing platforms handle and manage all of these data and requests. However, for certain applications, the data transmission delay that comes with transferring data from edge devices to the cloud could be unbearable. When there are a lot of devices connected to the internet, the public network actually becomes a bottleneck for data transfer. In this setting, power management, data storage, resource management, and service management all necessitate more robust infrastructure and complex processes. More efficient use of network and cloud resources is achievable with fog computing's “intelligent gateway” capability. Planning and managing resources is one of the most important factors affecting system performance (especially latency) in a fog-cloud environment. Planning in an environment with fog and clouds is an NP-hard problem. This paper delves into the optimisation difficulty of longevity for data-intensive job scheduling in fog and cloud-based IoT systems. The issue is initially expressed as an optimisation model for integer linear programming (ILP). Next, we provide a heuristic algorithm known as DLJSF (Data-Locality Aware Job Scheduling in Fog-Cloud) that is based on the suggested formulation. The results of the tests showed that the performance of the proposed algorithm is close to the results by an average of 87 %. Also, on average, it is 99.16 % better than the LP results obtained from the optimal solution obtained from the solver obtained from the solution that the data is processed locally. To check the efficiency of the simulation solution, it was repeated for tasks with different entry rates and data with different sizes. According to the obtained documents, the data transfer approach can be valuable and the proposed algorithm has not lost its performance in different conditions. •Presented here is a way for scheduling jobs in cloud-fog-based IoT systems that make use of data location knowledge by combining fog and cloud processing.•In order to reduce the overall quantity of makepan, the workplace in the fog system has been located through data replication and transmission between fogs, which goes against previous works.•Lessening the strain on public cloud resources and network bandwidth is the goal of the proposed method.