minimize Header

Prof. Dr. Andreas J. Kassler

Faculty of Computer Science

Professors

Professor

Subject: Networks and Distributed Systems

Mobile Networks (5G, 6G, WiFi)
Realtime Networks (TSN)
Programmable Netzworks
Application of Machine Learning in Networks
Intelligent Distributed Systems (e.g. Datacenter, SmartGrid,..)

DEGG's 2.19

0991/3615-8298

consulting time

upon request

publications

Sortierung:

Contribution

Andreas Fischer
D. Bhamare
Andreas Kassler

On the Construction of Optimal Embedding Problems for Delay-Sensitive Service Function Chains.

In: Proceedings of the 28th International Conference on Computer Communications and Networks (ICCCN 2019) [July 29-August 1, 2019; Valencia, Spain]. pg. 1-10

(2019)

Contribution

Y. Sharma
M. Khan
J. Taheri
Andreas Kassler

Performance Benchmarking of Virtualized Network Functions to Correlate Key Performance Metrics with System Activity.

In: 2020 11th International Conference on Network of the Future (NoF). pg. 73-81

IEEE

(2020)

DOI: 10.1109/NoF50125.2020.9249199

Journal article

M. Khoshkholghi
Michel Gokan K.
Kyoomars Alizadeh Noghani
J. Taheri
Deval Bhamare
Andreas Kassler
Zhengzhe Xiang
S. Deng
Xiaoxian Yang

Service Function Chain Placement for Joint Cost and Latency Optimization.

In: Mobile Networks and Applications vol. 25 pg. 2191-2205

(2020)

DOI: 10.1007/s11036-020-01661-w

Contribution

M. Khan
J. Taheri
M. Khoshkholghi
Andreas Kassler
C. Cartwright
M. Darula
S. Deng

A Performance Modelling Approach for SLA-Aware Resource Recommendation in Cloud Native Network Functions. Best Student Paper Award.

In: 2020 6th IEEE International Conference on Network Softwarization (NetSoft). pg. 292-300

IEEE

(2020)

DOI: 10.1109/NetSoft48620.2020.9165482

Journal article

J. Vestin
Andreas Kassler
S. Laki
G. Pongrácz

Toward In-Network Event Detection and Filtering for Publish/Subscribe Communication Using Programmable Data Planes.

In: IEEE Transactions on Network and Service Management vol. 18 pg. 415-428

(2021)

DOI: 10.1109/TNSM.2020.3040011

Contribution

S. Laki
P. Vörös
G. Szabo
Andreas Kassler

Revitalizing Industrial Networking with Programmable Data Planes.

In: 2021 P4 Workshop.

(2021)

Contribution

Phil Aupke
Andreas Kassler
Andreas Theocharis
Magnus Nilsson
Michael Uelschen

Quantifying Uncertainty for Predicting Renewable Energy Time Series Data Using Machine Learning.

In: The 7th International Conference on Time Series and Forecasting (ITISE) 2021. pg. 50

MDPI Basel, Switzerland

(2021)

DOI: 10.3390/engproc2021005050

Contribution

R. Figueiredo
Andreas Kassler

BNG-HAL: A Unified API for Disaggregated BNGs.

In: 2021 IEEE Conference on Network Function Virtualization and Software Defined Networks (NFV-SDN). pg. 116-119

IEEE

(2021)

DOI: 10.1109/NFV-SDN53031.2021.9665122

Journal article

N. Skorin-Kapov
R. Santos
H. Ghazzai
Andreas Kassler

A Randomized Greedy Heuristic for Steerable Wireless Backhaul Reconfiguration.

In: Electronics vol. 10 pg. 434

(2021)

DOI: 10.3390/electronics10040434

Contribution

C. Benet
Andreas Kassler
G. Antichi
T. Benson
G. Pongrácz

Providing In-network Support to Coflow Scheduling.

In: 2021 IEEE 7th International Conference on Network Softwarization (NetSoft). pg. 235-243

IEEE

(2021)

DOI: 10.1109/NetSoft51509.2021.9492530

Contribution

H. Chahed
Andreas Kassler

Software-Defined Time Sensitive Networks Configuration and Management.

In: 2021 IEEE Conference on Network Function Virtualization and Software Defined Networks (NFV-SDN). pg. 124-128

IEEE

(2021)

DOI: 10.1109/NFV-SDN53031.2021.9665120

Journal article

M. Gokan Khan
J. Taheri
A. Al-Dulaimy
Andreas Kassler

PerfSim: A Performance Simulator for Cloud Native Microservice Chains.

In: IEEE Transactions on Cloud Computing pg. 1-1

(2021)

DOI: 10.1109/TCC.2021.3135757

Contribution

A. Nammouchi
Phil Aupke
Andreas Kassler
Andreas Theocharis
Viviana Raffa
Marco Di Felice

Integration of AI, IoT and Edge-Computing for Smart Microgrid Energy Management.

In: 2021 IEEE International Conference on Environment and Electrical Engineering and 2021 IEEE Industrial and Commercial Power Systems Europe (EEEIC / I&CPS Europe). pg. 1-6

IEEE

(2021)

DOI: 10.1109/EEEIC/ICPSEurope51590.2021.9584756

Journal article

K. Noghani
Andreas Kassler
J. Taheri
P. Ohlen
C. Curescu

Multi-Objective Genetic Algorithm for Fast Service Function Chain Reconfiguration.

In: IEEE Transactions on Network and Service Management pg. 1-1

(2022)

DOI: 10.1109/TNSM.2022.3195820

Journal article

A. Mesodiakaki
E. Zola
Andreas Kassler

Robust and energy-efficient user association and traffic routing in B5G HetNets.

In: Computer Networks vol. 217 pg. 109305

(2022)

DOI: 10.1016/j.comnet.2022.109305

Journal article

A. Al-Dulaimy
J. Taheri
Andreas Kassler
M. Hoseiny Farahabady
S. Deng
Albert Zomaya

MultiScaler: A Multi-Loop Auto-Scaling Approach for Cloud-Based Applications.

In: IEEE Transactions on Cloud Computing vol. 10 pg. 2769-2786

(2022)

DOI: 10.1109/TCC.2020.3031676

Contribution

P. Aupke
Andreas Kassler
A. Theocharis
M. Nilsson
I. Myren Andersson

Impact of Clustering Methods on Machine Learning based Solar Power Prediction Models.

In: Proceedings of the IEEE International Smart Cities Conference (ISC2 2022).

(2022)

Contribution

R. Alfredsson
Andreas Kassler
J. Vestin
M. Pieska
M. Amend

Accelerating a Transport Layer Based 5G Multi-Access Proxy on SmartNIC.

In: Würzburg Workshop on Next-Generation Communication Networks - WueWoWas 2022.

(2022)

Contribution

Y. Ma
Andreas Kassler
B. Ahmed
P. Krakhmalev
A. Thore
A. Toyser
H. Lindbäck

Using Deep Reinforcement Learning for Zero Defect Smart Forging.

In: The 10th Swedish Production Symposium (SPSS) 2022.

(2022)

Contribution

M. Amend
N. Moreno
M. Pieska
Andreas Kassler
A. Brunstrom
V. Rakocevic
S. Johnson

In-network Support for Packet Reordering for Multiaccess Transport Layer Tunneling.

In: 2022 IEEE 11th IFIP International Conference on Performance Evaluation and Modeling in Wireless and Wired Networks (PEMWN). pg. 1-6

IEEE

(2022)

DOI: 10.23919/PEMWN56085.2022.9963814

Journal article

R. Santos
N. Skorin-Kapov
H. Ghazzai
Andreas Kassler
G. Tran

Towards the optimal orchestration of steerable mmWave backhaul reconfiguration.

In: Computer Networks vol. 205 pg. 108750

(2022)

DOI: 10.1016/j.comnet.2021.108750

Journal article

F. Bayram
B. Ahmed
Andreas Kassler

From concept drift to model degradation: An overview on performance-aware drift detectors.

In: Knowledge-Based Systems vol. 245 pg. 108632

(2022)

DOI: 10.1016/j.knosys.2022.108632

Journal article

S. Kumar Singh
C. Rothenberg
J. Langlet
Andreas Kassler
P. Voros
S. Laki
G. Pongrácz

Hybrid P4 Programmable Pipelines for 5G gNodeB and User Plane Functions.

In: IEEE Transactions on Mobile Computing pg. 1-18

(2022)

DOI: 10.1109/TMC.2022.3201512

Journal article

D. Bhamare
Andreas Kassler
J. Vestin
M. Khoshkholghi
J. Taheri
T. Mahmoodi
P. Öhlén
C. Curescu

IntOpt: In-band Network Telemetry optimization framework to monitor network slices using P4.

In: Computer Networks vol. 216 pg. 109214

(2022)

DOI: 10.1016/j.comnet.2022.109214

Contribution

A. Nammouchi
M. Stentani
S. Paoletti
Andreas Kassler
A. Theocharis

Robust Operation of Energy Communities in the Italian Incentive System.

In: Proceedings of the IEEE PES ISGT EUROPE 2023.

Eds.:
IEEE Power & Energy Society
Université Grenoble Alpes

(2023)

DOI: 10.1109/ISGTEUROPE56780.2023.10408430

Contribution

A. Nammouchi
Andreas Kassler
A. Theocharis

Quantum Machine Learning in Climate Change and Sustainability: a Short Review.

In: Proceedings of the AAAI 2023 Fall Symposium - Artificial Intelligence and Climate: The Role of AI in a Climate-Smart Sustainable Future.

(2023)

DOI: 10.1609/aaaiss.v2i1.27657

Contribution

R. Figueiredo
Andreas Kassler
H. Karl

Performance Measurements of Broadband Network Gateways.

In: Book of Abstracts of the Symposium on Electrical and Computer Engineering at the Doctoral Congress in Engineering (DCE) 2023. pg. 16-17

Eds.:
L. Almeida
T. Gonçalves
N. Fidalgo
G. Javanmardi

(2023)

show abstract

Broadband network access is typically managed by Broadband Network Gateways (BNGs), which implement complex services including authentication, authorization, and accounting (AAA), packet routing and forwarding, and Quality of Service (QoS) enforcement. QoS enforcement is especially complex to implement due to different subscriber tariffs, which require different traffic shaping, policing, and queue management policies to be supported in parallel at high fidelity. The recent network softwarization trend allows BNG functionality to be implemented as a Virtual Network Function (VNF). This enables flexible deployment strategies on commodity hardware, significantly reducing capital expenditure (CAPEX). However, the complex packet processing inside the BNG data plane makes it difficult to provide low and predictable latency at low loss at the scale required by access network subscribers. With the recent emergence of new, programmable, protocol-independent packet processing hardware targets (such as Field-Programmable Gate Arrays (FPGAs) or programmable ASICs), accelerating the BNG data plane on re-programmable hardware becomes feasible. But the vast range of deployment options makes it difficult to choose the programmable targets best suited for a particular key performance indicator. The architecture and services of the BNG are described in several technical reports of the Broadband Forum. The functional split between control and user planes of the BNG was defined by the TR-459 [1], additionally describing the interfaces between these two components. The in-depth description of the different services is found in the TR-178 [2]. The design of BNG functionality as a VNF has been as well the target of academic studies. In Kundel et al. [3], the BNG was implemented in P4-enabled hardware targets and shows the achievable performance for the different targets. Since P4 is not designed to support packet queueing and scheduling, FPGAs are used to realize QoS functionality. Mejia and Rothenberg [4] proposes a P4-based BNG, using the MACSAD as the execution environment. The design of QoS functionality as a VNF, in particular packet scheduling, has been studied as well. Fejes et al. [5] proposes a system capable of describing hierarchical scheduling policies without needing to maintain a large set of queues. Xi et al. [6] proposes the offload of the Linux hierarchical token bucket (HTB) to Netronome SmartNIC. In this work, we measure and compare the performance of the same functionality based on different implementations, to characterize the tradeoffs between performance and flexibility. To gain more insights into the different performance aspects of accelerating BNG packet processing functions, we perform a controlled benchmark study on the BNG use case on two targets. In particular, we first deploy a software version of BNG as a typical VNF on an x86 processor using the high-speed packet processing framework VPP. For the second BNG implementation, we disaggregate the data plane and implement typical BNG packet processing functions in P4 and deploy them on a programmable switching ASIC while traffic shaping is implemented on an FPGA. For the benchmark, we create scenarios representing different residential network access patterns, focussing on VoIP and IPTV services, which are sensitive to delay and loss. Therefore, we focus our evaluation on the performance of the BNG data plane, particularly on the enforcement of Quality-of-Service policies. We analyze the following key performance indicators: i) the throughput shaping accuracy for different policies; ii) the packet-processing delay and delay variation; iii) the energy consumption of the BNG data plane.

Journal article

A. Nammouchi
P. Aupke
F. D’Andreagiovanni
H. Ghazzai
A. Theocharis
Andreas Kassler

Robust opportunistic optimal energy management of a mixed microgrid under asymmetrical uncertainties.

In: Sustainable Energy, Grids and Networks vol. 36 pg. 101184

(2023)

DOI: 10.1016/j.segan.2023.101184

show abstract

Energy management within microgrids under the presence of large number of renewables such as photovoltaics is complicated due to uncertainties involved. Randomness in energy production and consumption make both the prediction and optimality of exchanges challenging. In this paper, we evaluate the impact of uncertainties on optimality of different robust energy exchange strategies. To address the problem, we present AIROBE, a data-driven system that uses machine-learning-based predictions of energy supply and demand as input to calculate robust energy exchange schedules using a multiband robust optimization approach to protect from deviations. AIROBE allows the decision maker to tradeoff robustness with stability of the system and energy costs. Our evaluation shows, how AIROBE can deal effectively with asymmetric deviations and how better prediction methods can reduce both the operational cost while at the same time may lead to increased operational stability of the system.

Journal article

H. Chahed
Andreas Kassler

TSN Network Scheduling—Challenges and Approaches.

In: Network vol. 3 pg. 585-624

(2023)

DOI: 10.3390/network3040026

show abstract

Time-Sensitive Networking (TSN) is a set of Ethernet standards aimed to improve determinism in packet delivery for converged networks. The main goal is to provide mechanisms that enable low and predictable transmission latency and high availability for demanding applications such as real-time audio/video streaming, automotive, and industrial control. To provide the required guarantees, TSN integrates different traffic shaping mechanisms including 802.1Qbv, 802.1Qch, and 802.1Qcr, allowing for the coexistence of different traffic classes with different priorities on the same network. Achieving the required quality of service (QoS) level needs proper selection and configuration of shaping mechanisms, which is difficult due to the diversity in the requirements of the coexisting streams under the presence of potential end-system-induced jitter. This paper discusses the suitability of the TSN traffic shaping mechanisms for the different traffic types, analyzes the TSN network configuration problem, i.e., finds the optimal path and shaper configurations for all TSN elements in the network to provide the required QoS, discusses the goals, constraints, and challenges of time-aware scheduling, and elaborates on the evaluation criteria of both the network-wide schedules and the scheduling algorithms that derive the configurations to present a common ground for comparison between the different approaches. Finally, we analyze the evolution of the scheduling task, identify shortcomings, and suggest future research directions.

Contribution

Y. Ma
K. Younis
B. Ahmed
Andreas Kassler
P. Krakhmalev
A. Thore
H. Lindbäck

Automated and Systematic Digital Twins Testing for Industrial Processes.

In: 2023 16th IEEE International Conference on Software Testing, Verification and Validation Workshops (ICST). pg. 149-158

IEEE

(2023)

DOI: 10.1109/ICSTW58534.2023.00037

Contribution

P. Aupke
Seema
A. Theocharis
Andreas Kassler
D.-E. Archer

PV Power Production and Consumption Estimation with Uncertainty bounds in Smart Energy Grids.

In: 2023 IEEE International Conference on Environment and Electrical Engineering and 2023 IEEE Industrial and Commercial Power Systems Europe (EEEIC / I&CPS Europe). pg. 1-6

IEEE

(2023)

DOI: 10.1109/EEEIC/ICPSEurope57605.2023.10194894

show abstract

For efficient energy exchanges in smart energy grids under the presence of renewables, predictions of energy production and consumption are required. For robust energy scheduling, prediction of uncertainty bounds of Photovoltaic (PV) power production and consumption is essential. In this paper, we apply several Machine Learning (ML) models that can predict the power generation of PV and consumption of households in a smart energy grid, while also assessing the uncertainty of their predictions by providing quantile values as uncertainty bounds. We evaluate our algorithms on a dataset from Swedish households having PV installations and battery storage. Our findings reveal that a Mean Absolute Error (MAE) of 16.12W for power production and 16.34W for consumption for a residential installation can be achieved with uncertainty bounds having quantile loss values below 5W. Furthermore, we show that the accuracy of the ML models can be affected by the characteristics of the household being studied. Different households may have different data distributions, which can cause prediction models to perform poorly when applied to untrained households. However, our study found that models built directly for individual homes, even when trained with smaller datasets, offer the best outcomes. This suggests that the development of personalized ML models may be a promising avenue for improving the accuracy of predictions in the future.

Journal article

F. Bayram
P. Aupke
B. Ahmed
Andreas Kassler
A. Theocharis
J. Forsman

DA-LSTM: A dynamic drift-adaptive learning framework for interval load forecasting with LSTM networks.

In: Engineering Applications of Artificial Intelligence vol. 123, Part C pg. 106480

(2023)

DOI: 10.1016/j.engappai.2023.106480

show abstract

Load forecasting is a crucial topic in energy management systems (EMS) due to its vital role in optimizing energy scheduling and enabling more flexible and intelligent power grid systems. As a result, these systems allow power utility companies to respond promptly to demands in the electricity market. Deep learning (DL) models have been commonly employed in load forecasting problems supported by adaptation mechanisms to cope with the changing pattern of consumption by customers, known as concept drift. A drift magnitude threshold should be defined to design change detection methods to identify drifts. While the drift magnitude in load forecasting problems can vary significantly over time, existing literature often assumes a fixed drift magnitude threshold, which should be dynamically adjusted rather than fixed during system evolution. To address this gap, in this paper, we propose a dynamic drift-adaptive Long Short-Term Memory (DA-LSTM) framework that can improve the performance of load forecasting models without requiring a drift threshold setting. We integrate several strategies into the framework based on active and passive adaptation approaches. To evaluate DA-LSTM in real-life settings, we thoroughly analyze the proposed framework and deploy it in a real-world problem through a cloud-based environment. Efficiency is evaluated in terms of the prediction performance of each approach and computational cost. The experiments show performance improvements on multiple evaluation metrics achieved by our framework compared to baseline methods from the literature. Finally, we present a trade-off analysis between prediction performance and computational costs.

Contribution

F. Brisch
Andreas Kassler
J. Vestin
M. Pieska
M. Amend

Accelerating Transport Layer Multipath Packet Scheduling for 5G-ATSSS.

In: Würzburg Workshop on Next-Generation Communication Networks - WueWoWas 2023.

(2023)

Journal article

H. Chahed
M. Usman
A. Chatterjee
F. Bayram
R. Chaudhary
A. Brunstrom
J. Taheri
B. Ahmed
Andreas Kassler

AIDA—A holistic AI-driven networking and processing framework for industrial IoT applications.

In: Internet of Things vol. 22 pg. 100805

(2023)

DOI: 10.1016/j.iot.2023.100805

show abstract

Industry 4.0 is characterized by digitalized production facilities, where a large volume of sensors collect a vast amount of data that is used to increase the sustainability of the production by e.g. optimizing process parameters, reducing machine downtime and material waste, and the like. However, making intelligent data-driven decisions under timeliness constraints requires the integration of time-sensitive networks with reliable data ingestion and processing infrastructure with plug-in support of Machine Learning (ML) pipelines. However, such integration is difficult due to the lack of frameworks that flexibly integrate and program the networking and computing infrastructures, while allowing ML pipelines to ingest the collected data and make trustworthy decisions in real time. In this paper, we present AIDA - a novel holistic AI-driven network and processing framework for reliable data-driven real-time industrial IoT applications. AIDA manages and configures Time-Sensitive networks (TSN) to enable real-time data ingestion into an observable AI-powered edge/cloud continuum. Pluggable and trustworthy ML components that make timely decisions for various industrial IoT applications and the infrastructure itself are an intrinsic part of AIDA. We introduce the AIDA architecture, demonstrate the building blocks of our framework and illustrate it with two use cases.

Lecture

Andreas Kassler
M. Amend

Explorative Journey on 5G and beyond – Dos and Don't’s. Interactive Session.

In: KuVS Fachgespräch - Würzburg Workshop on Next-Generation Communication Networks

Würzburg

28.-30.06.2023 (2023)

Lecture

F. Brisch
Andreas Kassler
J. Vestin
M. Pieska
M. Amend

Accelerating Transport Layer Multipath Packet Scheduling for 5G-ATSSS.

In: KuVS Fachgespräch - Würzburg Workshop on Next-Generation Communication Networks

Würzburg

28.-30.06.2023 (2023)

Contribution

M. Memarian
Andreas Kassler
K.-J. Grinnemo
S. Laki
G. Pongrácz
J. Forsman

Utilizing Hybrid P4 Solutions to Enhance 5G gNB with Data Plane Programmability.

In: Würzburg Workshop on Next-Generation Communication Networks – WueWoWas 2024.

(2024)

Contribution

R. Figueiredo
H. Woesner
Andreas Kassler
H. Karl

Quality of Service Performance of Multi-Core Broadband Network Gateways.

In: 2024 8th Network Traffic Measurement and Analysis Conference (TMA). pg. 1-10

IEEE

(2024)

DOI: 10.23919/TMA62044.2024.10559123

show abstract

Broadband network access is typically managed by Broadband Network Gateways (BNGs), which can be implemented as a Virtual Network Function (VNF). This paradigm shift is caused by network softwarization and allows the BNG to be deployed on commodity hardware, significantly reducing capital expenditure (CAPEX). But packet processing operations and complex Quality of Service (QoS) policies make it difficult to provide low and predictable latency at scale for a large number of subscribers. To improve performance, parallel queues at the Network Interface Card (NIC) and multiple dedicated CPU cores for packet processing are used, processing 50 million packets per second on commodity x86 hardware. How to guarantee latency, however, remains unclear. In this study, we conducted testbed-based experiments on a VPP/DPDK implementation of the BNG to benchmark its performance. Our findings reveal how latency and its variation increase with background traffic, and we analyze a parameter that contributes to a trade-off between throughput and latency. We also examine the ability of the multi-core architecture to guarantee latency, at a cost of reduced port utilization. These observations influence the design goal of isolating subscriber traffic and highlight the suitability of software BNG for guaranteeing performance.

Contribution

F. Brisch
Andreas Kassler
S. Laki
P. Hudoba

P4-MTAGG - a Framework for Multi-Tenant P4 Network Devices.

In: 2024 20th International Conference on Network and Service Management (CNSM). pg. 1-3

IEEE

(2024)

DOI: 10.23919/CNSM62983.2024.10814635

show abstract

The current P4 programmability model assumes that a P4 programmable device is owned and controlled by a single tenant. However, in typical NFV scenarios, support for multiple tenants is desirable. When each tenant may want to deploy their own P4 pipeline offering different network functions (NF), supporting multiple co-existing tenant pipelines on a single platform is difficult because it requires pipeline merging, control plane support, and resource management of the platform. In this paper, we present P4-MTAGG, a novel framework for flexibly deploying multiple P4 programmable NFs on a programmable match-action pipeline while supporting multiple tenants. P4-MTAGG consists of i) novel compiler-add-ons for automatic merging multiple P4-pipelines, ii) p4runtime-proxy to allow for control plane access of the aggregated pipelines together with policy-based resource management for the P4 target, and iii) orchestrator to automate the provisioning of a network node utilizing aggregation either in a simulated or real hardware environment. In this demo, we show how P4-MTAGG aggregates multiple NFs of varying complexity in Mininet. The user can orchestrate the aggregation process through a GUI. The per-tenant traffic is routed through the set of NFs using segment routing. Through the GUI, the user can instruct the p4runtime-proxy to enforce per-tenant bandwidth limits, which configure the per-tenant available resources in the data plane.

Contribution

H. Chahed
Andreas Kassler

Optimizing TSN Routing, Scheduling, and Task Placement in Virtualized Edge-Compute Platforms.

In: 2024 27th Conference on Innovation in Clouds, Internet and Networks (ICIN). pg. 153-157

IEEE

(2024)

DOI: 10.1109/ICIN60470.2024.10494455

show abstract

Configuring TSN network elements involves solving a (joint) routing and scheduling problem, where traditionally the TSN endpoints (i.e. talkers and listeners) are already deployed inside fixed industrial computers. However, with the emergence of edge computing on the shop floor, PLCs are becoming virtualized and can flexibly be deployed in containers or virtual machines. This additional flexibility could enhance the network configuration. In this paper, we propose GenTSN, a hybrid genetic algorithm designed to jointly optimize TSN routing, scheduling, and placement of TSN tasks (i.e. talkers and listeners) in virtualized Edge-Compute Platforms. We evaluate GenTSN, showing its efficiency compared to state of the art scheduling algorithms. In particular, we demonstrate that additional degrees of freedom to flexibly place TSN tasks or to flexibly route the traffic leads to better performance.

Contribution

P. Aupke
A. Nakao
Andreas Kassler

Uncertainty-Aware Forecasting of Computational Load in MECs Using Distributed Machine Learning: A Tokyo Case Study.

In: 2024 33rd International Conference on Computer Communications and Networks (ICCCN). pg. 1-6

IEEE

(2024)

DOI: 10.1109/ICCCN61486.2024.10637613

show abstract

Mobile Edge Clouds (MECs) address the critical needs of bandwidth-intensive, latency-sensitive mobile applications by positioning computing and storage resources at the network's edge in Edge Data Centers (EDCs). However, the diverse, dynamic nature of EDCs' resource capacities and user mobility poses significant challenges for resource allocation and management. Efficient EDC operation requires accurate forecasting of computational load to ensure optimal scaling, service placement, and migration within the MEC infrastructure. This task is complicated by the temporal and spatial fluctuations of computational load.We develop a novel MEC computational demand forecasting method using Federated Learning (FL). Our approach leverages FL's distributed processing to enhance data security and prediction accuracy within MEC infrastructure. By incorporating uncertainty bounds, we improve load scheduling robustness. Evaluations on a Tokyo dataset show significant improvements in forecast accuracy compared to traditional methods, with a 42.04% reduction in Mean Absolute Error (MAE) using LightGBM and a 34.93% improvement with CatBoost, while maintaining minimal networking overhead for model transmission.

Contribution

A. Nammouchi
C. Chaabani
A. Theocharis
Andreas Kassler

Towards Explainable Renewable Energy Communities Operations Using Generative AI.

In: IEEE PES ISGT Europe 2024 Conference Book. pg. 110

Eds.:
T. Baškarad
N. Holjevac
I. Kuzle
M. Zidar

(2024)

show abstract

Renewable Energy Communities (RECs), characterized by decentralized energy networks, are a key enabler for enhancing renewable energy utilization, cost-efficient planning and clean energy transition. However, optimizing RECs operations is challenging due to the complex interplay of different stakeholders with conflicting requirements. The complexity of managing such systems often leads to a lack of transparent and reliable decision-making, creating barriers for actors within the community. This paper explores the integration of Generative AI into Renewable Energy Communities (RECs) to enhance the transparency, explainability and accessibility of energy management systems (EMSs) that depend on solving optimization problems. We propose a novel framework, Chat-SGP, which uses generative AI to synthesize optimization modeling code to provide actionable, explainable insights for managing the REC operations. Our approach allows us to interact with the EMS through natural language queries, enhancing the system's accessibility and user- friendliness. Our evaluation shows that using GPT-4 with in-context learning performs 96.72\% accuracy on average in generating correct answers.

Contribution

H. Chahed
F. Hallström
H. Alcaine
Andreas Kassler

Linux-Based End-Station Design for Seamless TSN Plug -and - Play.

In: 2024 IEEE International Conference on Advanced Systems and Emergent Technologies (IC_ASET). pg. 1-6

IEEE

(2024)

DOI: 10.1109/IC_ASET61847.2024.10596212

show abstract

In TSN networks, proper end-station configuration is essential to ensure the timely and reliable delivery of time-sensitive data, meeting strict end- to-end Quality of Service (QoS) criteria. However, the complexity of the configuration process requires a significant manual effort, which makes real-time application development on standard Operating Systems such as Linux a challenge. In this paper, we propose a sim-ple yet functional approach to automate the configuration of Linux-based TSN end-stations within TSN networks by adding a TSN layer on top of the networking system services and defining a configuration protocol tailored for the centralized network/distributed user configuration mode. Evaluation results demonstrate minimal overhead during stream addition, achieving hundreds-of-millisecond-Ievel configuration times and enabling a hassle-free Plug-and-Play mode of operation.

Contribution

F. Brisch
Andreas Kassler
S. Laki
P. Hudoba
G. Pongrácz

P4-MTAGG - a Framework for Multi-Tenant P4 Network Devices.

In: Würzburg Workshop on Next-Generation Communication Networks – WueWoWas 2024.

(2024)

Journal article

M. Pieska
Andreas Kassler
A. Brunstrom
V. Rakocevic
M. Amend

Performance Impact of Nested Congestion Control on Transport-Layer Multipath Tunneling.

In: Future Internet vol. 16 pg. 233

(2024)

DOI: 10.3390/fi16070233

show abstract

Multipath wireless access aims to seamlessly aggregate multiple access networks to increase data rates and decrease latency. It is currently being standardized through the ATSSS architectural framework as part of the fifth-generation (5G) cellular networks. However, facilitating efficient multi-access communication in next-generation wireless networks poses several challenges due to the complex interplay between congestion control (CC) and packet scheduling. Given that enhanced ATSSS steering functions for traffic splitting advocate the utilization of multi-access tunnels using congestion-controlled multipath network protocols between user equipment and a proxy, addressing the issue of nested CC becomes imperative. In this paper, we evaluate the impact of such nested congestion control loops on throughput over multi-access tunnels using the recently introduced Multipath DCCP (MP-DCCP) tunneling framework. We evaluate different combinations of endpoint and tunnel CC algorithms, including BBR, BBRv2, CUBIC, and NewReno. Using the Cheapest Path First scheduler, we quantify and analyze the impact of the following on the performance of tunnel-based multipath: (1) the location of the multi-access proxy relative to the user; (2) the bottleneck buffer size, and (3) the choice of the congestion control algorithms. Furthermore, our findings demonstrate the superior performance of BBRv2 as a tunnel CC algorithm.

Contribution

M. Khan
J. Taheri
Andreas Kassler
A. Asl

Graph Attention Networks and Deep Q-Learning for Service Mesh Optimization: A Digital Twinning Approach.

In: ICC 2024 - IEEE International Conference on Communications. pg. 2913-2918

IEEE

(2024)

DOI: 10.1109/ICC51166.2024.10622616

show abstract

In the realm of cloud native environments, Ku-bernetes has emerged as the de facto orchestration system for containers, and the service mesh architecture, with its interconnected microservices, has become increasingly prominent. Efficient scheduling and resource allocation for these microservices play a pivotal role in achieving high performance and maintaining system reliability. In this paper, we introduce a novel approach for container scheduling within Kubernetes clusters, leveraging Graph Attention Networks (GATs) for representation learning. Our proposed method captures the intricate dependencies among containers and services by constructing a representation graph. The deep Q-learning algorithm is then employed to optimize scheduling decisions, focusing on container-to-node placements, CPU request-response allocation, and adherence to node affinity and anti-affinity rules. Our experiments demonstrate that our GATs-based method outperforms traditional scheduling strategies, leading to enhanced resource utilization, reduced service latency, and improved overall system throughput. The insights gleaned from this study pave the way for a new frontier in cloud native performance optimization and offer tangible benefits to industries adopting microservice-based architectures.

Journal article

M. Pieska
A. Rabitsch
A. Brunstrom
Andreas Kassler
M. Amend
E. Bogenfeld

Low-delay cost-aware multipath scheduling over dynamic links for access traffic steering, switching, and splitting.

In: Computer Networks vol. 241 pg. 110186

(2024)

DOI: 10.1016/j.comnet.2024.110186

show abstract

Bundling of multiple access technologies is currently being standardized by 3GPP in the 5G access traffic steering, switching and splitting (ATSSS) framework, with the goal to increase robustness, resiliency and capacity of wireless access. A key part of an ATSSS framework is the packet scheduler, which decides the access network over which each packet is to be transmitted. As wireless channels are highly dynamic, a challenge for any scheduler is to correctly estimate the capacity of each path, and thereby avoid congesting the paths. In this paper, we further develop a recent packet scheduler that exploits cross-layer information from the congestion control state of individual transport layer tunnels when making scheduling decisions. Our aim is to achieve good path utilization while keeping the congestion delay low. Extensive emulations show that our approach reduces the excess delay at the bottleneck to as little as 34%. We furthermore show that our approach improves the performance of end-to-end applications including WebRTC and YouTube compared to state-of-the art.

Contribution

M. Memarian
Andreas Kassler
K.-J. Grinnemo
S. Laki
G. Pongrácz
J. Forsman

Utilizing Hybrid P4 Solutions to Enhance 5G gNB with Data Plane Programmability.

In: 15th Joint IFIP Wireless and Mobile Networking Conference (WMNC).

(2024)

Book

Özgür Kaynak
Andreas Kassler
Andreas Fischer
O. Dobrijevic
H. Chahed

TSN Scheduling Robust to Wireless Performance Uncertainties: A Problem and Model Definition.

Universität Würzburg Würzburg

(2025)

show abstract

Time-sensitive Networking (TSN) is a set of extensions to the Ethernet standard for providing deterministic communication services over a converged network infrastructure. A key element thereby is meeting end-to-end latency requirements, which can be achieved for wired links by properly configuring TSN endpoints and switches using IEEE 802.1AS time synchronization and an IEEE 802.1Qbv time-aware shaper. However, wireless links commonly exhibit performance uncertainties, which introduce additional challenges for deterministic communication. This paper proposes a new method for configuring TSN networks with wireless links, focusing on cyclically scheduled traffic. We design a linear program that synthesizes TSN configurations robust to performance uncertainties of wireless links, by adjusting the transmission schedule at the first wired network node. Moreover, we develop an opportunistic version of the scheduling when the senders transmit multiple frames within their sending interval.

projects

Spitzenprofessur der HighTech Agenda

labs

Intelligent Networks and Systems (Head)

core competencies

Software Defined Networking
Programmable Dataplanes
Network (Function) Virtualization
Autonomic Networking
Future Internet
Wireless Networks
Network Optimization

Forschungs- und Lehrgebiete

Andreas J. Kassler is Professor of Computer Science at Deggendorf Institute of Technology, Germany (since 2023) and Karlstads Universitet, Karlstad, Sweden (since 2005). From 2003 to 2004, Dr. Andreas J. Kassler was Assistant Professor at the School of Computer Engineering, Nanyang Technological University, Singapore. At Degegndorf, he is leading the Intelligent Network and Systems Lab. He maintains an active research program in the fields of networking and cloud computing with main research focus on Software Defined Networking, Future Internet, Datacenter Networking and, Quality of Service.

Dr. Andreas J. Kassler received the Docent title (Habilitation) in Computer Science from Karlstads Universitet in 2007 and the Ph.D. degree in Computer Science from Universität Ulm, Germany, in 2002. He received the M. Sc. degree in Mathematics/Computer Science in 1995 from Universität Augsburg, Germany.

He is co-author of around 55 peer reviewed journal articles and book chapters, 195 peer reviewed conference and workshop publications, 7 European or international patents and 11 IETF and ISO standardization contributions. He is also co-editor of a book published in the LNCS book series of Springer. He is the area editor of the Elsevier Computer Networks Journal, served as a guest editor of a feature topic in EURASIP Wireless Communications and Networking Journal, and served as Associate Editor on the editorial boards of some refereed international journals, such as: Journal of Internet Engineering, International Journal On Advances in Networks and Services.