.png){kind=icon}
Return
An efficient function placement approach in serverless edge computing
.png){kind=icon}
PRE
AI
DOI：10.1007/s00607-025-01438-7.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Serverless computing has gained significant attention due to its promise of simplifying application development and deployment. Application providers in this computing model must implement their applications using primarily stateless functions, and they do not need complex infrastructure management. Due to the ever-increasing expansion of IoT devices and real-time services, serverless computing has become popular at the edge. IoT devices use many applications in serverless edge computing. In serverless edge computing, we face requests with different requirements and workloads for executing functions that must be placed and executed on heterogeneous edge devices in such a way that they meet the user's requirements and the quality of service. This problem is known as dynamic function placement in serverless edge computing, and it is one of the critical challenges in this computing. In this paper, we introduce an autonomous dynamic function placement approach using the autonomic computing model and the deep reinforcement learning technique to make decisions about dynamic deploying functions in heterogeneous and dynamic edge infrastructure. An autonomous function placement framework is also designed based on the three-layer architecture of the public edge environment. When comparing the proposed solution with the other methods, simulation results indicate that the proposed solution reduces average Cost by 27.6% and delays by 28.8% while increasing edge node utilization by 18.6%.
Keywords:
Serverless edge computing
Function placement
Autonomic computing
Deep reinforcement learning
Journal
C
IF:
2.8
Papers:
2.3K
Citations:
3.5K
Organization
Cited Papers
.png){kind=icon}
ShareSaveRemote patient monitoring using artificial intelligence: Current state, applications, and challenges
OAAI
OAAIShareSave
ShareSaveShareSave
View more