Denne rapporten er utarbeidet av Center for Resilient Networks and Applications (CRNA), som er en del av Simula Metropolitan Center for Digital Engineering. CRNA driver grunn- leggende forskning innen robusthet og sikker- het i nettverk med mandat og finansiering fra Kommunal- og moderniseringsdepartementet. Senteret produserer en årlig rapport om tilstan- den i norske mobilnett. Årets rapport er den niende i rekken.

This white paper on AI and ML as enablers of beyond 5G (B5G) networks is based on contributions from almost 20 5G PPP projects, coordinated through the 5G PPP Technology Board, that research, implement and validate 5G and B5G network systems. The paper introduces the main relevant mechanisms in Artificial Intelligence (AI) and Machine Learning (ML), currently investigated and exploited for enhancing 5G and B5G networks.

With the increasing deployment of the Multi-Path Transmission Control Protocol (MPTCP) in heterogeneous network setups, there is a need to understand how its performance is affected in practice both by traditional factors such as round-trip time measurements, buffer predictive modelling and by calculating the impact factors of network subflows. Studies have shown that path management and packet scheduling have a large effect on overall performance and required limited resources with different congestion control parameters. Unfortunately, most of the previous studies have focused almost exclusively on the improvement of a single parameter, without a holistic view. To deal with this issue effectively, this paper puts forward a Multi-Parameter Comprehensive Optimized Algorithm (MPCOA), which can find the smaller buffer size and select the appropriate congestion control and path management algorithm on the premise of ensuring larger throughput. Experiments of three scenarios show that MPCOA can save the buffer space and subflow resources, and achieve high throughput. Meanwhile, a set of quantitative improvement results given by MPCOA is convenient for us to evaluate the quality of MPTCP network, and provide reference for our ongoing future work, like for 4G/5G, Internet of Things and Star Link networks.

A large fraction of the communication in the Internet is handled by the Transmission Control Protocol (TCP). Since the first deployments of this protocol more than 30 years ago, the spectrum of applications as well as the structure of the network have developed at a fast pace. For example, today's network devices, like smartphones and laptops — i.e. particularly many devices in the area of mobile computing — frequently have an interesting property: the existence of multiple IP addresses (IPv4 and/or IPv6). The addresses may even change due to mobility. This property, denoted as multi-homing, can be utilised for multi-path transport, i.e. the simultaneous usage of multiple paths in the network to improve performance. Multi-path transport is a hot topic in the Internet Engineering Task Force (IETF), which is the standardisation organisation for the Internet. This talk provides an overview of the work in the areas of multi-homing and multi-path transport, with focus on the area of the protocols TCP and Stream Control Transmission Protocol (SCTP) with their experimental extensions Multi-Path TCP (MPTCP) and Concurrent Multi-Path Transfer for SCTP (CMT-SCTP). It particularly shows the sequence of research and selected results, beginning from a simple simulation model, via lab setups and small Internet scenarios, up to the large-scale, international testbed project NorNet. NorNet, and particularly its landline network part NorNet Core, is furthermore described in some detail. Based on NorNet, it is finally possible to validate simulation results in real-world, multi-homed networks, in order to provide valuable input to the ongoing IETF standardisation processes of MPTCP and CMT-SCTP. Particularly, it will also show how the NorNet testbed can be utilised for research at Hainan University in 2021.

The SimulaMet OpenAirInterface VNF provides an OpenAirInterface-based Enhanced Packet Core (EPC), with separate VDUs for HSS, MME, SPGW-C and SPGW-U. To allow for advanced in-band telemetry, we have extended this VNF to add switches with Programming Protocol-independent Packet Processors (P4) to all relevant virtual links inside the EPC. P4 allows full programability of the packet forwarding behaviour, and especially allows to extend packets with additional information for in-band telemetry. This information can be read by other P4 instances to allow for fine-granular performance data collection. In this presentation and live demonstration, we would like show the solutions chosen to efficiently use OSM for handling our extended EPC, and in particular we would like to highlight the possibilities to perform P4-based in-band telemetry to evaluate the performance of the mobile network. Finally, we would also like to show the audience a live demo of our testbed setup with telemetry collection.

This tutorial — presented for students at the College of Information Science and Technology (CIST) at Hainan University — provides an introduction on how to get access to the NorNet Core testbed as well as how to run experiments in the testbed in 2021.

Network APIs are moving towards protocol agility, where applications express their needs but not a static protocol binding, and it is up to the layer below the API to choose a suitable protocol. The IETF Transport Services (TAPS) Working Group is standardizing a protocol-independent transport API and offering guidance to implementers. Apple’s recent “Network.framework” is specifically designed to allow such late and dynamic binding of protocols. When the network stack autonomously chooses and configures a protocol, it must first test which protocols are locally available and which work end-to-end (“protocol racing”). For this, it is important to know the set of available options, and which protocols should be tried first: Does it make sense to offer unchecked payload delivery, as with UDP-Lite? Is a UDP-based protocol like QUIC always a better choice, or should native SCTP be tried? This paper develops answers to such questions via (i) a NAT study in a local testbed, (ii) bidirectional Internet tests, (iii) a large scale Internet measurement campaign. The examined protocols are: SCTP, DCCP, UDP-Lite, UDP with a zero checksum and three different UDP encapsulations.