Ákos Leiter, Döme Matusovits, and László Bokor
Evaluation of traditional and eBPF-based packet processing in Kubernetes for network slicing
In recent years, the proliferation of cloud-native technology enablers, such as microservice deployment and management with Kubernetes, have presented new challenges for telecommunications service providers. Strict data transmission requirements have emerged in various areas, such as immediate interventions in intelligent transportation, video conferencing, etc. With the advent of 5G networks, this demand can also be fulfilled thanks to an innovative technology called Network Slicing. In terms of its operation, we can separate networks into individual segments to continuously satisfy the desired service requirements. However, packet processing on top of Kubernetes may need to be changed to support the emerging number of microservices during slicing. This is where the Extended Berkeley Packet Filter (eBPF) comes into the picture to boost the capacity of data centers and keep service guarantees. This paper presents how eBPF can support network slicing through its performance evaluation in a Kubernetes environment.
Reference:
DOI: 10.36244/ICJ.2025.1.7
Please cite this paper the following way:
Ákos Leiter, Döme Matusovits, and László Bokor, "Evaluation of traditional and eBPF-based packet processing in Kubernetes for network slicing", Infocommunications Journal, Vol. XVII, No 1, March 2025, pp. 56-65., https://doi.org/10.36244/ICJ.2025.1.7