.png){kind=icon}
.png){kind=icon}
Return
Network service function chaining: a performance study varying topologies
.png){kind=icon}
PRE
AI
DOI：10.1007/s10586-024-04778-9.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
NFV (Network Function Virtualization) is an innovative paradigm that seeks to improve the flexibility and efficiency of network services. NFV focuses on virtualizing network functions, allowing them to be executed on generic hardware, which decouples software and hardware and allows for more agile and scalable implementations. Through NFV, the concept of SFC (Service Function Chain) emerges, which consists of a sequence of VNFs (Virtual Network Functions), such as gateways, firewalls, and routers, arranged in an orderly manner. A critical challenge faced by SFC is its sensitivity to latency: delay increases as the chain of VNFs lengthens, which can compromise the quality of service. Minimizing this latency is crucial to ensure the efficiency of SFC systems. However, testing different SFC architectures in real environments is costly. As a solution, this study proposes using Stochastic Petri Nets (SPNs) for a performance evaluation of SFC architectures. The developed SPN model allows for analyzing different configurations of VNFs in sequential, parallel, and hybrid architectures. Four case studies were carried out exploring the behavior of the proposed model under different analyses and scenarios. The results obtained were validated through simulations in GNS3 (Graphical Network Simulator-3). The data demonstrate that changes in the SFC architecture can result in significant performance improvements and latency reduction.
Keywords:
NFV
SFC
VNF
Performance
Petri nets
Journal
C
IF:
4.1
Papers:
5.0K
Citations:
7.5K