Scholarlink

<p>Due to the significant increase in network traffic demands, traditional cloud systems are increasingly concerned about maintaining Quality of Service (QoS) and Quality of Experience (QoE). This challenge led to the introduction of fog/edge computing. However, the resource constraints of these environments make designing effective resource management critical, requiring greater attention, particularly through caching strategies that store frequently accessed content to enhance QoS and QoE. Accurately predicting content popularity patterns plays a crucial role in optimizing caching performance, especially under dynamic user demands. The demonstrated capability of Deep Learning (DL) architectures in modeling complex temporal-spatial patterns has consequently established them as strong solutions for next-generation caching systems. This review paper provides an overview of fog/edge computing environments and their challenges. We then concentrate on the caching issue within these environments, proposing caching taxonomies from various points of view that have not been mentioned by previous researchers. Additionally, we reviewed some studies that utilize DL techniques to address caching challenges. We compare these studies based on caching challenges, the DL techniques employed, and their main features. In conclusion, this study covers challenges and limitations that need more investigation to improve both fog/edge computing and content caching strategies. This survey also gives future directions, which are so useful for researchers to find their way and can help them to solve open challenges and limitations in these environments.</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 Survey on Edge/Fog Caching Based on Deep Learning Algorithms

  • 期刊论文

  • 发表日期：

    2025年10月9日

查看全文
  • Bahareh Shafaie,
  • Marjan Kuchaki Rafsanjani,
  • Varsha Arya,
  • Brij B. Gupta

摘要

Due to the significant increase in network traffic demands, traditional cloud systems are increasingly concerned about maintaining Quality of Service (QoS) and Quality of Experience (QoE). This challenge led to the introduction of fog/edge computing. However, the resource constraints of these environments make designing effective resource management critical, requiring greater attention, particularly through caching strategies that store frequently accessed content to enhance QoS and QoE. Accurately predicting content popularity patterns plays a crucial role in optimizing caching performance, especially under dynamic user demands. The demonstrated capability of Deep Learning (DL) architectures in modeling complex temporal-spatial patterns has consequently established them as strong solutions for next-generation caching systems. This review paper provides an overview of fog/edge computing environments and their challenges. We then concentrate on the caching issue within these environments, proposing caching taxonomies from various points of view that have not been mentioned by previous researchers. Additionally, we reviewed some studies that utilize DL techniques to address caching challenges. We compare these studies based on caching challenges, the DL techniques employed, and their main features. In conclusion, this study covers challenges and limitations that need more investigation to improve both fog/edge computing and content caching strategies. This survey also gives future directions, which are so useful for researchers to find their way and can help them to solve open challenges and limitations in these environments.