Learning SDN traffic flow accurate models to enable queue bandwidth dynamic optimization

Reticcioli, Enrico; Girolamo, Giovanni Domenico Di; Smarra, Francesco; Carmenini, A.; D’Innocenzo, Alessandro; Graziosi, Fabio

doi:10.1109/eucnc48522.2020.9200916

Cited by 5 publications

(4 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The traffic demand matrix and current network states are provided for learning agent training, and a Deep Q-network (DQN) determines the optimal feasible path. Reticcioli et al [21] proposed an AutoregRessive eXogenous model (ARX) that enables the accurate queue bandwidth management of SDN switches. The method combines a regression tree and random forest and can dynamically control the switchport queue input/output behavior of network devices.…”

Section: Qos Support In Single-domain Sdnmentioning

confidence: 99%

See 1 more Smart Citation

QoS Support Path Selection for Inter-Domain Flows Using Effective Delay and Directed Acyclic Graph in Multi-Domain SDN

Lee

Roh

2022

Electronics

View full text Add to dashboard Cite

Currently, network applications, such as audio, video, and augmented reality, have different stringent service requirements. They require service provision through end-to-end connections via other networks with different operating environments or service conditions. Therefore, network operators require information on their own and other networks to provide end-to-end services traversing several networks while guaranteeing their quality of service (QoS) requirements. This study proposes an inter-domain flow decision method that satisfies QoS requirements using a directed acyclic graph (DAG) in multi-domain and hierarchical software defined networking (SDN) networks. There are multiple local networks with SDN controllers that are connected to the global SDN controller. The flow decision in the proposed method is based on the effective bandwidth theory of the martingale process. The effectiveness of the proposed method is demonstrated by comparing it with existing SDN-based path selection methods using the Riverbed Modeler (older, OPNET) and OpenDaylight SDN controllers.

show abstract

Section: Qos Support In Single-domain Sdnmentioning

confidence: 99%

“…As the number of terminals and the various types of data served through the network have increased, QoS support is also an essential issue in the management of SDN networks. Several studies have been conducted to provide QoS in single-domain SDN networks [5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20][21].…”

Section: Introductionmentioning

confidence: 99%

QoS Support Path Selection for Inter-Domain Flows Using Effective Delay and Directed Acyclic Graph in Multi-Domain SDN

Lee

Roh

2022

Electronics

View full text Add to dashboard Cite

show abstract

“…[11] and references therein). In particular, some of these estimation techniques have been successfully applied to experimental use‐cases in energy efficient building automation [12], in structural health monitoring and control [13] and in control of Software Defined communication Networks [14]. While the learning approaches leveraged by the aforementioned works have exhibited impressive experimental results, they tend to produce unstable models even if the original system is stable.…”

Section: Introductionmentioning

confidence: 99%

Stability conditions for linear discrete‐time switched systems in block companion form

Iuliis

D’Innocenzo

Germani

et al. 2020

IET Control Theory & Appl

View full text Add to dashboard Cite

Switched models whose dynamic matrices are in block companion form arise in theoretical and applicative problems such as representing switched ARX models in state-space form for control purposes. Inspired by some insightful results on the delay-independent stability of discrete-time systems with time-varying delays, in this work, the authors study the arbitrary switching stability for some classes of block companion discrete-time switched systems. They start from the special case in which the first block-row is made of permutations of non-negative matrices, deriving a simple necessary and sufficient stability condition under arbitrary switching. The condition is computationally less demanding than the sufficient-only existence of a linear common Lyapunov function. Then, both non-negativity and combinatorial assumptions are dropped, at the expense of introducing conservatism. Some implications on the computation of the joint spectral radius for the aforementioned families of matrices are illustrated.

show abstract

“…We also show that our methodology quickly improves its performance as soon as new data are available, bringing to an improvement of both model accuracy and control performance. The present work is based on the preliminary conference paper in [30], which has been improved and extended in the following aspects: (i) the effect of iterative model updates using on-the-fly new data in the prediction accuracy is demonstrated; (ii) the effect of predictive models of future incoming traffic is tested; (iii) the accuracy of the predictive models is validated over a real dataset obtained from network measurements of an Italian Internet Service Provider (Sonicatel S.r.l. ).…”

mentioning

confidence: 99%

Modeling and Control of Priority Queueing in Software Defined Networks via Machine Learning

et al. 2022

View full text Add to dashboard Cite

Software Defined Networking (SDN) is a new architectural paradigm that enables programmable control of a network to make it more flexible and easier to manage. SDN architectures decouple control and forwarding functionalities, and enable switches and routers to be remotely configurable/programmable in run-time by a controller. Modeling and optimization of such modern heterogeneous network infrastructures are key factors to achieve better performance, e.g. in terms of traffic flow improvement while reducing bandwidth allocation. Identifying an accurate model of a network device in SDNs (e.g., a switch or a router) is crucial in order to apply advanced techniques such as Model Predictive Control (MPC). However, such a problem is very challenging due to non-linearities and unavailability of internal variables measurements in real devices. To this aim, a promising direction is given by an appropriate integration of System Identification and Machine Learning techniques to obtain predictive models using historical data collected from the network thanks to the SDN paradigm. In this paper we apply a novel data-driven methodology to learn accurate models of the dynamical input-output behavior of a network's switch device by appropriately combining AutoRegressive eXogenous (ARX) model identification with Regression Trees (RTs) and Random Forests (RFs). The advantage of such model is that it can be directly used to apply MPC (which just requires Quadratic Programming to be solved) to optimally control the queues' bandwidth of the switch ports within the SDN paradigm. We validate our approach on an experimental emulation setup using the Mininet network emulator environment and a real dataset obtained from measurements of an Italian Internet Service Provider (Sonicatel S.r.l.). To this aim, we first develop a model of a real network switch, then implement MPC using the RYU controller, and finally demonstrate the benefits of the proposed dynamic queueing control methodology in terms of packet losses reduction and bandwidth saving, i.e. in terms of improvement of the Quality of Service.

show abstract

Learning SDN traffic flow accurate models to enable queue bandwidth dynamic optimization

Cited by 5 publications

References 22 publications

QoS Support Path Selection for Inter-Domain Flows Using Effective Delay and Directed Acyclic Graph in Multi-Domain SDN

QoS Support Path Selection for Inter-Domain Flows Using Effective Delay and Directed Acyclic Graph in Multi-Domain SDN

Stability conditions for linear discrete‐time switched systems in block companion form

Modeling and Control of Priority Queueing in Software Defined Networks via Machine Learning

Contact Info

Product

Resources

About