Software-based networking and processing

Accelerator technologies are conquering the domain of network and AI processors in the form of software accelerator mechanisms such as DPDK or eBPF, or as hardware components like netFPGAs, smartNICs, GPUs, TPUs, or NPUs. They offer tremendous potential to improve the performance and capabilities of networks or their supported services in compute continuum involving heterogeneous IoT/Edge and cloud systems.

The performance trade-offs of these novel platforms however are currently not yet fully understood and introduce plenty of uncertainties on how these platforms perform under which conditions. IDLab performs research in performance profiling, resource allocation and orchestration algorithms on top of heterogeneous hardware involving accelerator options. IDLab has an excellent track record in this domain and can build further on extensive expertise in SDN and NFV control mechanisms built in the context of national and international research projects.

Copyright © 2024 IDLab. All rights reserved.