Routing centralization across domains via SDN: A model and emulation framework for BGP evolution

Kotronis, Vasileios; Adrian, Gamperli,; Dimitropoulos, Xenofontas

doi:10.1016/j.comnet.2015.07.015

Cited by 31 publications

(37 citation statements)

References 52 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Due to the flexibility and programmability of the SDN, there are potentially many ways in which routing can be conducted. In this section, we describe a simple routing path construction mechanism which is generalized based on principles of BGP-like protocol [6] in the Internet, and routing mechanisms employed by some state-of-the-art controllers such as the ONOS controller. Note that it is not our intention to design any routing mechanisms; we use this simple and representative mechanism for the sake of problem formulation.…”

Section: A Generalized Path Construction Mechanism In Sdnmentioning

confidence: 99%

DQ Scheduler: Deep Reinforcement Learning Based Controller Synchronization in Distributed SDN

Zhang

Ma²,

Poularakis

et al. 2019

ICC 2019 - 2019 IEEE International Conference on Communications (ICC)

View full text Add to dashboard Cite

In distributed software-defined networks (SDN), multiple physical SDN controllers, each managing a network domain, are implemented to balance centralized control, scalability and reliability requirements. In such networking paradigm, controllers synchronize with each other to maintain a logically centralized network view. Despite various proposals of distributed SDN controller architectures, most existing works only assume that such logically centralized network view can be achieved with some synchronization designs, but the question of how exactly controllers should synchronize with each other to maximize the benefits of synchronization under the eventual consistency assumptions is largely overlooked. To this end, we formulate the controller synchronization problem as a Markov Decision Process (MDP) and apply reinforcement learning techniques combined with deep neural network to train a smart controller synchronization policy, which we call the Deep-Q (DQ) Scheduler. Evaluation results show that DQ Scheduler outperforms the antientropy algorithm implemented in the ONOS controller by up to 95.2% for inter-domain routing tasks.

show abstract

Section: A Generalized Path Construction Mechanism In Sdnmentioning

confidence: 99%

DQ Scheduler: Deep Reinforcement Learning Based Controller Synchronization in Distributed SDN

Zhang

Ma²,

Poularakis

et al. 2019

ICC 2019 - 2019 IEEE International Conference on Communications (ICC)

View full text Add to dashboard Cite

show abstract

“…The proposed solutions have demonstrated that bringing SDN to inter-domain routing can indeed improve the convergence performance of BGP [12], offer new routing capabilities [6], or lay the groundwork for new services and markets [13], [7]. However, most of previous works are systemoriented: they propose new systems or architectures, and focus on design or implementation aspects.…”

Section: Introductionmentioning

confidence: 99%

Can SDN accelerate BGP convergence?— A performance analysis of inter-domain routing centralization

Sermpezis

Dimitropoulos

2017

2017 IFIP Networking Conference (IFIP Networking) and Workshops

Self Cite

View full text Add to dashboard Cite

The Internet is composed of Autonomous Systems (ASes) or domains, i.e., networks belonging to different administrative entities. Routing between domains/ASes is realised in a distributed way, over the Border Gateway Protocol (BGP). Despite its global adoption, BGP has several shortcomings, like slow convergence after routing changes, which can cause packet losses and interrupt communication even for several minutes. To accelerate convergence, inter-domain routing centralization approaches, based on Software Defined Networking (SDN), have been recently proposed. Initial studies show that these approaches can significantly improve performance and routing control over BGP. In this paper, we complement existing system-oriented works, by analytically studying the gains of inter-domain SDN. We propose a probabilistic framework to analyse the effects of centralization on the inter-domain routing performance. We derive bounds for the time needed to establish data plane connectivity between ASes after a routing change, as well as predictions for the control-plane convergence time. Our results provide useful insights (e.g., related to the penetration of SDN in the Internet) that can facilitate future research. We discuss applications of our results, and demonstrate the gains through simulations on the Internet AS-topology. • We propose a model (Section II) and methodology (Sections III and IV) for the performance analysis of interdomain routing centralization. To our best knowledge, we are the first to employ a probabilistic approach to study the performance of inter-domain SDN. • We analyse the time that the network needs to establish connectivity after a routing change. In particular, we derive upper and lower bounds for the time needed to achieve data-plane connectivity between two ASes (Section III), and exact expressions and approximations for the time till control-plane convergence over the entire network (Section IV). Our results are given by closed-form expressions, as a function of network parameters, like network size, path lengths, and number of SDN nodes. • Based on the theoretical expressions, as well as on extensive simulation results, we provide insights for potential gains of centralization, inter-domain SDN deployment strategies, network economics, etc.We believe that our study can be useful in a number of directions. Research in inter-domain SDN can be accelerated

show abstract

“…However, when it comes to hijacking of /24 prefixes, the de-aggregation might not be always efficient, since prefixes more specific than /24 are filtered by most routers [25]. Although this is a shortcoming of the de-aggregation mechanism, it is not possible to overcome it in an automatic way (manual actions are needed); to our best knowledge, only solutions that require the cooperation of more than one ASes could be applied [26,27].…”

Section: Automatic Prefix Hijacking Mitigationmentioning

confidence: 99%

Artemis

Chaviaras¹,

Gigis²,

Sermpezis³

et al. 2016

Proceedings of the 2016 ACM SIGCOMM Conference

Self Cite

View full text Add to dashboard Cite

The Border Gateway Protocol (BGP) is globally used by Autonomous Systems (ASes) to establish route paths for IP prefixes in the Internet. Due to the lack of authentication in BGP, an AS can hijack IP prefixes owned by other ASes (i.e., announce illegitimate route paths), impacting thus the Internet routing system and economy. To this end, a number of hijacking detection systems have been proposed. However, existing systems are usually third party services that -inherently-introduce a significant delay between the hijacking detection (by the service) and its mitigation (by the network administrators). To overcome this shortcoming, in this paper, we propose ARTEMIS, a tool that enables an AS to timely detect hijacks on its own prefixes, and automatically proceed to mitigation actions. To evaluate the performance of ARTEMIS, we conduct real hijacking experiments. To our best knowledge, it is the first time that a hijacking detection/mitigation system is evaluated through extensive experiments in the real Internet. Our results (a) show that ARTEMIS can detect (mitigate) a hijack within a few seconds (minutes) after it has been launched, and (b) demonstrate the efficiency of the different control-plane sources used by ARTEMIS, towards monitoring routing changes.

show abstract

Routing centralization across domains via SDN: A model and emulation framework for BGP evolution

Cited by 31 publications

References 52 publications

DQ Scheduler: Deep Reinforcement Learning Based Controller Synchronization in Distributed SDN

DQ Scheduler: Deep Reinforcement Learning Based Controller Synchronization in Distributed SDN

Can SDN accelerate BGP convergence?— A performance analysis of inter-domain routing centralization

Artemis

Contact Info

Product

Resources

About