Knowledge-Defined Networking

摘要

I start by applauding the authors of Knowledge-Defined Networking for pub­licly sharing the datasets and scripts used in the experiments presented in the paper, at inednetworking.org . This is not only im­portant for reproducibility, but for the particular context of Machine Learn­ing (ML) having standardised datasets has been shown fundamental in other areas, as the authors correctly point out, so this is an excellent first step towards that goal in networking. The website provides two main types of artefacts: datasets and neural net­works software. The datasets include all data used for the two use cases discussed in the paper. For the virtual network functions they include the CPU consumption of an OVS connected to an SDN controller, of an OVS configured with firewall rules, and of SNORT. For the network characterisa­tion use case, the authors include several delay measurements for different network topologies. The datasets have shown to be good and useful.