Year: 2019
Type: conference-paper
Abstract: Network Function Virtualization aims to migrate packet-processing tasks from special-purpose appliances to Virtual Network Functions (VNFs) running on x86 or ARM servers. However, achieving the line-rate packet-processing for VNFs running on CPUs can be a challenging task. External hardware accelerators can be used to offload heavy-lifting tasks (e.g. en/decryption and hashing) from performance-critical VNFs.State-of-the-art VNF placement algorithms only consider compute resources while assigning VNFs on server nodes. We propose a placement algorithm which takes into consideration hardware accelerator resources in addition to compute resources. For evaluation, we compare the performance of our approach with the Integer Linear Program (ILP) method and also with the state-of-the-art best-fit method of VNF placement.
Institutions: Ghent University, IMEC
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
Funder: European Commission
OpenAlex ID: W3012012908
