Design of Osmotic Computing Based Task Offloading and Scheduling Algorithms in Healthcare Systems

Abstract

newline Healthcare has become an on-the-fly service available at one s fingertips via smartphones and wearables, enabling efficient management of many tasks. However, computationally intensive tasks lead to high energy consumption, and the digital transformation of healthcare generates vast data, causing latency issues. Immediate actions for latency-critical tasks require edge or cloud offloading, offering a solution to energy and latency challenges. newlineThis thesis introduces osmotic computing-based task mapping, task offloading, and task scheduling algorithms for healthcare systems. Osmotic computing is a neoteric paradigm that integrates cloud, edge, and Internet of Things (IoT) platforms. However, managing the eclectic range of computational resources across federated environment is tedious to manage. newlineWe present three task offloading algorithms, namely energy and latency-balanced osmotic-offloading (ELBO), fuzzy-osmotic-based task offloading (FOTO) and emergency level-based task offloading (ELTO). ELBO balances energy and latency by offloading tasks between IoT/edge and edge/cloud environments, enhancing performance. FOTO employs fuzzy logic to optimize offloading, assigning latency-sensitive tasks to IoT/edge and latency- tolerant tasks to the cloud. ELTO categorizes tasks by urgency as critical or normal, offloading them to edge or cloud accordingly. newlineAdditionally, we present an osmotic task scheduling (OTS) algorithm that dynamically assigns tasks to the most suitable resources based on computational requirements. The task mapping algorithm categorizes incoming tasks by resource needs and optimizes mapping decisions through task-level and processing-level categorization. newlineThe proposed algorithms are simulated and evaluated using synthetic test cases, demonstrating efficacy in metrics such as average latency, success ratio, task failure rate, energy consumption, and resource utilization. newlineKeywords: Osmotic Computing; Task Mapping; Task Offloading; Task Scheduling; Healthcare Systems; Health Monitori