Cluster Computing, Journal Year: 2023, Volume and Issue: 27(4), P. 4621 - 4633
Published: Nov. 30, 2023
Language: Английский
Internet of Things, Journal Year: 2023, Volume and Issue: 23, P. 100866 - 100866
Published: July 1, 2023
Language: Английский
Citations
59
Future Generation Computer Systems, Journal Year: 2023, Volume and Issue: 152, P. 55 - 69
Published: Oct. 28, 2023
Edge/fog computing, as a distributed computing paradigm, satisfies the low-latency requirements of ever-increasing number IoT applications and has become mainstream paradigm behind applications. However, because large require execution on edge/fog resources, servers may be overloaded. Hence, it disrupt also negatively affect applications' response time. Moreover, many are composed dependent components incurring extra constraints for their execution. Besides, environments inherently dynamic stochastic. Thus, efficient adaptive scheduling in heterogeneous is paramount importance. limited computational resources imposes an burden applying optimal but computationally demanding techniques. To overcome these challenges, we propose Deep Reinforcement Learning-based application Scheduling algorithm, called DRLIS to adaptively efficiently optimize time balance load servers. We implemented practical scheduler FogBus2 function-as-a-service framework creating edge-fog-cloud integrated serverless environment. Results obtained from extensive experiments show that significantly reduces cost by up 55%, 37%, 50% terms balancing, time, weighted cost, respectively, compared with metaheuristic algorithms other reinforcement learning
Language: Английский
Citations
40
IEEE Access, Journal Year: 2023, Volume and Issue: 11, P. 88339 - 88358
Published: Jan. 1, 2023
Microservices architecture is a new trend embraced by many organizations as way to modernize their legacy applications. However, although there work related the migration process, gap in body of knowledge principles they should adopt when implementing microservices architecture. This presents comprehensive survey, gathering literature that explores fundamental underlying object-oriented approach and how these concepts are monolithic architectures. In addition, our research encompasses both architectures microservices, along with an investigation into design patterns utilized within microservices. Our contribution present list used architecture, comparation between expounded experts decomposition architectures, Martin Fowler Sam Neuman, forerunner Principle Information Hiding, David Parnas, who discusses modularization mechanism improve flexibility understanding system. Additionally, we expose advantages disadvantages obtained from review carried out summary form, which can help reference for researchers academia industry finally reveal trends today.
Language: Английский
Citations
31
IEEE Access, Journal Year: 2024, Volume and Issue: 12, P. 118007 - 118059
Published: Jan. 1, 2024
Language: Английский
Citations
10
Smart Cities, Journal Year: 2024, Volume and Issue: 7(1), P. 680 - 711
Published: Feb. 19, 2024
The management of decentralized energy resources and smart grids needs novel data-driven low-latency applications services to improve resilience responsiveness ensure closer real-time control. However, the large-scale integration Internet Things (IoT) devices has led generation significant amounts data at edge grid, posing challenges for traditional cloud-based smart-grid architectures meet stringent latency response time requirements emerging applications. In this paper, we delve into grid computational distribution architectures, including edge–fog–cloud models, orchestration, frameworks support design offloading across continuum. Key factors influencing process, such as network performance, Artificial Intelligence (AI) processes, requirements, application-specific factors, efficiency, are analyzed considering operational requirements. We conduct a comprehensive overview current research landscape decision-making regarding strategies from cloud fog or edge. focus is on metaheuristics identifying near-optimal solutions reinforcement learning adaptively optimizing process. A macro perspective determining when what offload in provided next-generation AI applications, offering an features trade-offs selecting between federated solutions. Finally, work contributes understanding grids, providing Strengths, Weaknesses, Opportunities, Threats (SWOT) analysis cost–benefit strategies.
Language: Английский
Citations
8
Multimedia Tools and Applications, Journal Year: 2023, Volume and Issue: 83(8), P. 23019 - 23045
Published: Aug. 10, 2023
Language: Английский
Citations
17
Future Generation Computer Systems, Journal Year: 2024, Volume and Issue: 154, P. 479 - 490
Published: Jan. 21, 2024
Language: Английский
Citations
7
IEEE Transactions on Services Computing, Journal Year: 2024, Volume and Issue: 17(5), P. 1983 - 1997
Published: May 10, 2024
Microservices along with cloud-edge computing technologies are widely adopted to take advantage of the abundant resources cloud and low latency, high bandwidth capabilities edge. However, factors such as frequent user requirement changes have made current deployment scheme not fully adaptable new requirements, resulting in an increase average response time. Therefore, microservice system needs adjust online continuously changing requirements reduce time, which is known collaborative problem. existing methods able meet efficiency do consider complex dependencies microservices budget constraints cloud. To address this problem, paper proposes a solution problem by modeling it NP-hard integer nonlinear programming environment consisting private edge clouds public An efficient Two-stage Iterated Greedy Optimization (TIGO) algorithm also proposed its convergence proven. Extensive experimental results show that approach achieves better times less time compared methods.
Language: Английский
Citations
5
IEEE Access, Journal Year: 2023, Volume and Issue: 11, P. 60915 - 60938
Published: Jan. 1, 2023
The fog and edge computing paradigm provide a distributed architecture of nodes with processing capability for smart healthcare systems driven by Internet Thing (IoT) applications. It also provides method to reduce big data transmissions that cause latency enhance the system's efficiency. Resource provisioning scheduling in is significant problem due heterogeneity dispersion edge/fog/cloud resources. goal map tasks appropriate resources, which belong NP-hard problems, it takes much time find an optimal solution. Meta-heuristic methods achieve near-optimal solutions within reasonable time. Current edge/fog resource allocation research does not sufficiently address problems mobility-aware microservice-based IoT This paper proposes meta-heuristic-based micro-service model mobility management systems. proposed approach has been tested on experimental set-up simulation critical real-time application without considering devices. applies meta-heuristic such as modified genetic flower pollination algorithms management. outperforms existing energy consumption, network usage, cost, execution time, 17%, 20%, 22%, 63%, respectively.
Language: Английский
Citations
11
Future Generation Computer Systems, Journal Year: 2024, Volume and Issue: 157, P. 543 - 557
Published: April 8, 2024
Language: Английский
Citations
4