Adaptable and Data-Driven Softwarized Networks: Review, Opportunities, and Challenges

Kellerer, Wolfgang; Kalmbach, Patrick; Blenk, Andreas; Basta, Arsany; Reisslein, Martin; Schmid, Stefan

doi:10.1109/jproc.2019.2895553

Cited by 98 publications

(47 citation statements)

References 95 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…3GPP Release 15 features V2X using the emerging 5G mobile internet [30] which improves on LTE-V in every aspect. 5G is expected to achieve latencies as short as one millisecond and throughput of up 10 Gbit/s for up to 100 billion independent devices as well as greatly increased capacity [31]- [34]. 5G in combination with emerging lowlatency multi-access edge computing [35]- [37] appears thus well suited to enable intricate real-time control decisions in V2X scenarios.…”

Section: B Vehicle To Infrastructure Communicationmentioning

confidence: 99%

Optimised Traffic Light Management Through Reinforcement Learning: Traffic State Agnostic Agent vs. Holistic Agent With Current V2I Traffic State Knowledge

Busch

Latzko

Reisslein

et al. 2020

IEEE Open J. Intell. Transp. Syst.

Self Cite

View full text Add to dashboard Cite

Traffic light control falls into two main categories: Agnostic systems that do not exploit knowledge of the current traffic state, e.g., the positions and velocities of vehicles approaching intersections, and holistic systems that exploit knowledge of the current traffic state. Emerging fifth generation (5G) wireless networks enable Vehicle-to-Infrastructure (V2I) communication to reliably and quickly collect the current traffic state. However, to the best of our knowledge, the optimized traffic light management without and with current traffic state information has not been compared in detail. This study fills this gap in the literature by designing representative Deep Reinforcement Learning (DRL) agents that learn the control of multiple traffic lights without and with current traffic state information. Our agnostic agent considers mainly the current phase of all traffic lights and the expired times since the last change. In addition, our holistic agent considers the positions and velocities of the vehicles approaching the intersections. We compare the agnostic and holistic agents for simulated traffic scenarios, including a road network from Barcelona, Spain. We find that the holistic system substantially increases average vehicle velocities and flow rates, while reducing CO2 emissions, average wait and trip times, as well as a driver stress metric. Index Terms-Deep reinforcement learning (DRL), intelligent transportation system (ITS), intersection control, vehicle-toinfrastructure communication (V2I).

show abstract

Section: B Vehicle To Infrastructure Communicationmentioning

confidence: 99%

Optimised Traffic Light Management Through Reinforcement Learning: Traffic State Agnostic Agent vs. Holistic Agent With Current V2I Traffic State Knowledge

Busch

Latzko

Reisslein

et al. 2020

IEEE Open J. Intell. Transp. Syst.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Motivated by the complex, manual, and error-prone operation of today's communication networks, as well as the increasing dependability requirements in terms of availability and performance, the network community is currently very much engaged in developing more automated approaches to manage and operate networks. A particularly interesting vision in this context are self-driving networks [10,17]: rather than aiming for specific optimizations for certain protocols and objectives, networks should learn to drive themselves, maximizing high-level goals (such as end-to-end latency), in a "context-aware", data-driven manner. At the heart of such self-driving networks hence lies the ability to adaptively measure, analyze, and control themselves.…”

Section: Introductionmentioning

confidence: 99%

NetBOA

Zerwas

Kalmbach

Henkel

et al. 2019

Proceedings of the 2019 Workshop on Network Meets AI &Amp; ML - NetAI'19

Self Cite

View full text Add to dashboard Cite

Communication networks have not only become a critical infrastructure of our digital society, but are also increasingly complex and hence error-prone. This has recently motivated the study of more automated and "self-driving" networks: networks which measure, analyze, and control themselves in an adaptive manner, reacting to changes in the environment. In particular, such networks hence require a mechanism to recognize potential performance issues. This paper presents NetBOA, an adaptive and "data-driven" approach to measure network performance, allowing the network to identify bottlenecks and to perform automated what-if analysis, exploring improved network configurations. As a case study, we demonstrate how the NetBOA approach can be used to benchmark a popular software switch, Open vSwitch. We report on our implementation and evaluation, and show that NetBOA can find performance issues efficiently, compared to a non-data-driven approach. Our results hence indicate that NetBOA may also be useful to identify algorithmic complexity attacks.

show abstract

“…Such approaches neither work live and ad-hoc (i.e., without a prior information base) nor will they perform efficiently in the future due to the encryption of network traffic and increasing network sizes. However, efficient pattern extraction is a requirement to enable data driven algorithms that unlock the full potential of Software Defined Networking and Network Function Virtualization [10], as well as novel technologies such as re-configurable physical layers [8].…”

Section: Introductionmentioning

confidence: 99%

NOracle

Kalmbach

Hock²,

Lipp³

et al. 2019

Proceedings of the ACM SIGCOMM 2019 Conference Posters and Demos

Self Cite

View full text Add to dashboard Cite

This demo presents NOracle: a system using Stochastic Block Models (SBMs) to infer structural roles of hosts and communication patterns of services in networks. NOracle can be used with existing monitoring systems to analyze and visualize networks in an online manner or be used to analyze stored traces. Network operators can use SBMs to monitor and verify network operation, detect possible security issues and change-points. To showcase this, NOracle combines the production-grade network management solution StableNet with an SBM based anomaly detection and network visualization module. StableNet provides network flow statistics in real-time from actual devices. The SBM extracts roles and communication patterns live from the data provided by StableNet. The result can help to reason about communication behaviors, detect anomalous hosts and indicate changes in the large scale-structure of network communication.

show abstract

Adaptable and Data-Driven Softwarized Networks: Review, Opportunities, and Challenges

Abstract: This article reviews and analyzes adaptation opportunities and the potential for building scalable communication systems by means of software-defined networking (SDN) and network function virtualization (NFV).

Cited by 98 publications

References 95 publications

Optimised Traffic Light Management Through Reinforcement Learning: Traffic State Agnostic Agent vs. Holistic Agent With Current V2I Traffic State Knowledge

Optimised Traffic Light Management Through Reinforcement Learning: Traffic State Agnostic Agent vs. Holistic Agent With Current V2I Traffic State Knowledge

NetBOA

NOracle

Contact Info

Product

Resources

About