Scholarlink

<p>Cloud computing is a growing platform in which people are becoming increasingly interested. The services it offers provide robust and scalable features delivered through Information Technology and distributed over the Internet, which is an advantage over the usually higher fixed costs of in-house infrastructure and traditional systems management. As the Internet of Things (IoT) emerges, cloud-based systems have taken on even greater importance. However, there is no formal approach to model the risks in information systems implemented in a cloud environment. Service composition methods are important in cloud-based information systems as they help optimize resource allocation and increase service quality. In terms of data storage, processing power, and data analysis, IoT networks have almost unlimited potential in cloud systems. Cloud and IoT technologies are maturing together and coming together to create a cloud-enhanced IoT ecosystem today. Growing attention is needed to select and integrate individual services into unified solutions to satisfy client demands. However, the success of these methods is difficult to evaluate, as many metrics, such as Quality of Service (QoS), computational efficiency, and response time, are to be considered. This study puts forward a deep learning paradigm for service composition. This paradigm combines the capabilities of Long Short-Term Memory (LSTM) networks with Artificial Bee Colony (ABC) optimization. The LSTM component is used to make accurate Cloud QoS provision metric predictions, the output of which the ABC algorithm uses to find the optimal Cloud provider combinations for service composition, which minimizes end-users’ costs. The results obtained from the proposed method, compared to other algorithms showed the suggested algorithm’s appropriate performance in terms of availability, costs, response time, energy consumption, and reliability. Experimental results show that, compared to existing algorithms such as ABC, PSO, and GA, the proposed method reduced energy consumption by 18.2%, improved availability by 9.5%, shortened response time by 15.4%, and lowered operational cost by 12.8%.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A new efficient model to enhance the success of cloud-based service composition systems using an enhanced learning-based algorithm

  • 期刊论文

  • 发表日期：

    2025年10月17日

查看全文
  • Xijing Zhang,
  • ZhongJie Shen,
  • Rui Xie

摘要

Cloud computing is a growing platform in which people are becoming increasingly interested. The services it offers provide robust and scalable features delivered through Information Technology and distributed over the Internet, which is an advantage over the usually higher fixed costs of in-house infrastructure and traditional systems management. As the Internet of Things (IoT) emerges, cloud-based systems have taken on even greater importance. However, there is no formal approach to model the risks in information systems implemented in a cloud environment. Service composition methods are important in cloud-based information systems as they help optimize resource allocation and increase service quality. In terms of data storage, processing power, and data analysis, IoT networks have almost unlimited potential in cloud systems. Cloud and IoT technologies are maturing together and coming together to create a cloud-enhanced IoT ecosystem today. Growing attention is needed to select and integrate individual services into unified solutions to satisfy client demands. However, the success of these methods is difficult to evaluate, as many metrics, such as Quality of Service (QoS), computational efficiency, and response time, are to be considered. This study puts forward a deep learning paradigm for service composition. This paradigm combines the capabilities of Long Short-Term Memory (LSTM) networks with Artificial Bee Colony (ABC) optimization. The LSTM component is used to make accurate Cloud QoS provision metric predictions, the output of which the ABC algorithm uses to find the optimal Cloud provider combinations for service composition, which minimizes end-users’ costs. The results obtained from the proposed method, compared to other algorithms showed the suggested algorithm’s appropriate performance in terms of availability, costs, response time, energy consumption, and reliability. Experimental results show that, compared to existing algorithms such as ABC, PSO, and GA, the proposed method reduced energy consumption by 18.2%, improved availability by 9.5%, shortened response time by 15.4%, and lowered operational cost by 12.8%.