Elsevier

Computer Communications

Volume 77, 1 March 2016, Pages 41-51
Computer Communications

On reliable controller placements in Software-Defined Networks

https://doi.org/10.1016/j.comcom.2015.09.008Get rights and content

Abstract

In order to deploy fault-tolerant Software-Defined Networks (SDNs), a logically centralized controller must be physically distributed among different devices. In this paper, we focus on determining how many controllers need to be instantiated, where they must be deployed, and what network nodes are under control of each of them, in order to achieve high reliability in the southbound interface between controllers and nodes. For this, we define the fault tolerant controller placement problem and develop a heuristic algorithm that computes placements with (at least) the required reliability. We run such algorithm on a set of 124 publicly available network topologies. The results are thoroughly analyzed and provide insight on the feasibility of achieving fault tolerant SDNs by carefully determining the placement of controllers.

Section snippets

Introduction and motivation

Software-Defined Networks (SDNs) provide a clear separation between the control and data planes [1]. While traditional network nodes (switches, routers, middleboxes) integrate both planes within the same device, the SDN paradigm proposes the use of central controllers that dictate the behavior of the various forwarding elements (nodes) in the network. Control applications take advantage of the controller by querying and modifying the network state through a northbound interface. Similarly, the

Related work

The physical distribution of the control plane in SDN has been addressed in different works. Levin et al. explore the state distribution trade-offs between strongly consistent and eventually consistent models [7], given that network state is logically centralized but physically distributed among different controllers. While strongly consistent models ensure that all controllers perceive the same network view, they limit responsiveness in the control plane during the process of achieving

Problem formulation

Let G(V=NF;E) be the graph that represents the topology of a network, where N is the set of network nodes, F consists of the different facility locations where a controller can be deployed, and E corresponds to the links among them. Each vertex vV has a given probability pv of being operational (up and running). Analogously, links (u, v) ∈ E are operational with probability pu, v. We assume different i.i.d. operational probabilities for links, nodes, and controllers.

Let us define binary

Heuristic algorithm

Given the complexity of FTCP, we develop a heuristic algorithm that computes subsets Ij, ∀jN that meet the reliability constraint R^β. Since R^ is a lower bound, the solutions found by our algorithm provide an upper bound on the number of controllers required for β-reliability in the southbound interface.

The objective function of the FTCP optimization problem depends on the total number of controllers that are instantiated and on the number of controllers to which every node connects.

Analysis of fault-tolerant placements

Our interest in developing Algorithm 1 lies on its ability to analyze existing network topologies from the viewpoint of deploying fault-tolerant SDNs. Given a topology, if network nodes implement an SDN southbound protocol (e.g. OpenFlow), there is a set of facilities where controllers can be deployed, and we require that each node is effectively connected to at least one controller with high reliability, we want to answer the following questions: How many controllers must be deployed at least?

Discussion and conclusion

When nodes get disconnected from controllers, SDN applications lose visibility and control of part of the network. Therefore, in order to exploit the full potential of the SDN paradigm, it is important to design reliable southbound interfaces between nodes and controllers. Our work in this paper addresses such concern.

We evaluated solutions to the fault tolerant controller placement problem in a wide range of network topologies. Such solutions provide SDN operators with guidance on how many

Acknowledgment

This work has been partially supported by the European Commission through the Multi-Gigabit European Research and Education Network and Associated Services (GN3plus) project under Grant agreement no. 605243.

References (25)

  • Open Networking Foundation

    Software-defined networking: the new norm for networks

    ONF White Paper

    (2012)
  • T. Koponen et al.

    Onix: a distributed control platform for large-scale production networks

    Proceedings of the 9th USENIX Conference on Operating Systems Design and Implementation, OSDI’10

    (2010)
  • The OpenDaylight Project

    OpenDaylight – an open source community and meritocracy for software-defined networking

    White Paper

    (2013)
  • Open Networking Foundation

    OpenFlow Switch Specification. Version 1.2.0

    (2011)
  • B. Heller et al.

    The controller placement problem

    Proceedings of the ACM SIGCOMM Workshop on Hot Topics in Software Defined Networks, HotSDN’12

    (2012)
  • F. Ros et al.

    Five nines of southbound reliability in software-defined networks

    Proceedings of the ACM SIGCOMM Workshop on Hot Topics in Software Defined Networking, HotSDN

    (2014)
  • D. Levin et al.

    Logically centralized? State distribution trade-offs in software defined networks

    Proceedings of the ACM SIGCOMM Workshop on Hot Topics in Software Defined Networks, HotSDN’12

    (2012)
  • A. Tootoonchian et al.

    On controller performance in software-defined networks

    Proceedings of the Second USENIX Conference on Hot Topics in Management of Internet, Cloud, and Enterprise Networks and Services, Hot-ICE’12

    (2012)
  • S.H. Yeganeh et al.

    Kandoo: a framework for efficient and scalable offloading of control applications

    Proceedings of the ACM SIGCOMM Workshop on Hot Topics in Software Defined Networks, HotSDN’12

    (2012)
  • A. Tootoonchian et al.

    HyperFlow: a distributed control plane for OpenFlow

    Proceedings of the Internet Network Management Conference on Research on Enterprise Networking, INM/WREN’10

    (2010)
  • A. Dixit et al.

    Towards an elastic distributed SDN controller

    Proceedings of the ACM SIGCOMM Workshop on Hot Topics in Software Defined Networks, HotSDN’13

    (2013)
  • G. Yao et al.

    On the capacitated controller placement problem in software defined networks

    IEEE Commun. Lett.

    (2014)
  • Cited by (92)

    • A two-stage controller balanced deployment strategy based on inter-domain efficiency priority

      2023, Xi Tong Gong Cheng Yu Dian Zi Ji Shu/Systems Engineering and Electronics
    View all citing articles on Scopus
    View full text