NET2: Network Softwarisation II
Thursday, 10 June 2021, 16:00-17:30, Zoom Room
Session Chair: Thomas Henderson (Univ. Washington, USA)
Anomaly Detection and Analysis Framework for Mobile Networks
Jessica Mendoza, Isabel de-la-Bandera and Jesús Burgueño (University of Málaga, Spain); César Morillas and David Palacios (Tupl Inc., Spain); Raquel Barco (University of Malaga, Spain)
Proper management of failures in mobile communication networks is essential to provide quality services to users. This management consists of several tasks, being the first of them the detection of network failures. To carry out this task, key performance indicators (KPIs) that reflect the network state are analyzed. However, due to the different nature of these KPIs, the same detection method is not able to correctly find the anomalies in all of them. In addition, most of the techniques proposed at the moment, focus on the detection of certain types of anomalies. This paper proposes a framework for the detection of anomalies, capable of finding different types of anomalies in KPIs of different nature. This framework includes as well certain configuration parameters that allow to perform the detection based on the policies of network operators. As a result, the proposed framework indicates which of the anomalies found are actually KPI degradations as well as the start and end time of each degradation, and its percentage of degradation with respect to the normal behavior of the KPI.
Adaptive and Latency-Aware Load Balancing for Control Plane Traffic in the 4G/5G Core
Van Giang Nguyen, Karl-Johan Grinnemo, Javid Taheri and Anna Brunstrom (Karlstad University, Sweden)
For many years, the continuous proliferation of mobile devices and their applications generate a surge of signaling traffic in the control plane of the mobile packet core network. As a consequence, the control plane will potentially become a bottleneck if not properly managed. We focus on the load balancing of a virtualized and distributed Mobility Management Entity (vMME), which is the key control plane element in the 4G and 5G non-standalone cores. Most of existing works use the simple and static load balancing approaches such as round-robin and consistent hashing, which do not work well in a heterogeneous environment. In this paper, we developed three adaptive algorithms in which the balancing decision takes into account the dynamics of the system such as the vMME load, the completion time of a request served by a vMME, and the number of pending requests queued at a vMME. The evaluation of our three proposed load-balancing algorithms in an Open5GCore testbed suggests that the latency-aware scheme helps shorten the completion time of the signaling requests by up to five times the static and dynamic schemes in those cases the link delay between the load balancer and the vMMEs differ significantly.
On a Deep Q-Network-Based Approach for Active Queue Management
Dhulfiqar A AAlwahab (Eötvös Loránd University, Hungary); Gergo Gombos (ELTE Eötvös Loránd University, Hungary); Sándor Laki (Eötvös Loránd University, Hungary)
Reinforcement learning has gone through an enormous evolution in the past ten years. It’s practical applicability has been demonstrated through several use cases in various fields from robotics to process automation. In this paper, we examine how the tools of deep Q-learning can be used in an AQM algorithm to reduce queuing delay and ensure good link utilization at the same time. The proposed method called RL-AQM has the advantage that it is less prone to the good parameterization and can automatically adapt to new network conditions. The prototype implementation based on OpenAI Gym and NS-3 network simulator has thoroughly been evaluated under various settings focusing on three aspects: the convergence time of learning process, the performance of pre-trained models compared to PIE AQM and generalization ability. We have demonstrated that RL-AQM achieves comparable utilization to PIE AQM but results in much smaller queueing delays. Finally, the pre-trained models have good generalization abilities, enabling to use a pre-trained model in network settings that differ in bandwidth and/or RTT from the one used during the pre-training phase.
TeraFlow: Secured Autonomic Traffic Management for a Tera of SDN Flows
Ricard Vilalta, Raul Muñoz, Ramon Casellas and Ricardo Martinez (Centre Tecnològic de Telecomunicacions de Catalunya (CTTC/CERCA), Spain); Victor Lopez (Telefonica, Spain); Oscar González de Dios (Telefonica I+D, Spain); Antonio Pastor (Telefonica I+D & Universidad Politécnica de Madrid, Spain); Georgios Katsikas (Ubitech, Greece); Felix Klaedtke (NEC Europe Ltd., Germany); Paolo Monti (Chalmers University of Technology, Sweden); Alberto Mozo (Universidad Politécnica de Madrid, Spain); Thomas Zinner (NTNU, Norway); Harald Øverby (Norwegian University of Science and Technology, Norway); Sergio González (Atos, Spain); Håkon Lønsethagen (Telenor Research, Norway); José-Miguel Pulido (Volta Networks, Spain); Daniel King (Old Dog Consulting, United Kingdom (Great Britain))
TeraFlow proposes a new type of secure, cloud-native Software Defined Networking (SDN) controller that will radically advance the state-of-the-art in beyond 5G networks by introducing novel micro-services architecture, and provide revolutionary features for both flow management (service layer) and optical/microwave network equipment integration (infrastructure layer) by adapting new data models. TeraFlow will also incorporate security using Machine Learning (ML) and forensic evidence for multi-tenancy based on Distributed Ledgers. Finally, this new SDN controller shall be able to integrate with the current Network Function Virtualization (NFV) and Multi-access Edge Computing (MEC) frameworks as well as to other networks. The target pool of TeraFlow stakeholders expands beyond the traditional telecom operators towards edge and hyperscale cloud providers.
5Growth Data-Driven AI-Based Scaling
Danny De Vleeschauwer (Nokia, Belgium); Jorge Baranda (Centre Tecnològic de Telecomunicacions de Catalunya (CTTC/CERCA), Spain); Josep Mangues-Bafalluy (Centre Tecnològic de Telecomunicacions de Catalunya (CTTC), Spain); Carla Fabiana Chiasserini, Marco Malinverno and Corrado Puligheddu (Politecnico di Torino, Italy); Lina Magoula (National and Kapodistrian University of Athens, Greece); Jorge Martín-Pérez (Universidad Carlos III de Madrid, Spain); Sokratis Barmpounakis (University of Athens, Greece); Koteswararao Kondepu (Indian Institute of Technology Dharwad, India); Luca Valcarenghi (Scuola Superiore Sant’Anna, Italy); Xi Li (NEC, Germany); Chrysa Papagianni (University of Amsterdam, The Netherlands); Andres Garcia-Saavedra (NEC Labs Europe, Germany)
This paper presents a data-driven approach leveraging AI/ML models to automate the service scaling operation and, in this way, meet the service requirements while minimizing the consumption of network, computing, and storage resources. This approach is integrated into the 5Growth service management software platform. In particular, a prototype was developed to demonstrate how the novel 5Growth AI/ML platform can be used in a closed-loop automation system to support the automated service scaling operation. Furthermore, a number of additional ML-based approaches are developed in the context of eMBB and V2N scenarios, which can be embedded into the system for handling more complex use cases.