DISCO: Distributed Multi-domain SDN Controllers (1308.6138v2)

Abstract: Modern multi-domain networks now span over datacenter networks, enterprise networks, customer sites and mobile entities. Such networks are critical and, thus, must be resilient, scalable and easily extensible. The emergence of Software-Defined Networking (SDN) protocols, which enables to decouple the data plane from the control plane and dynamically program the network, opens up new ways to architect such networks. In this paper, we propose DISCO, an open and extensible DIstributed SDN COntrol plane able to cope with the distributed and heterogeneous nature of modern overlay networks and wide area networks. DISCO controllers manage their own network domain and communicate with each others to provide end-to-end network services. This communication is based on a unique lightweight and highly manageable control channel used by agents to self-adaptively share aggregated network-wide information. We implemented DISCO on top of the Floodlight OpenFlow controller and the AMQP protocol. We demonstrated how DISCO's control plane dynamically adapts to heterogeneous network topologies while being resilient enough to survive to disruptions and attacks and providing classic functionalities such as end-point migration and network-wide traffic engineering. The experimentation results we present are organized around three use cases: inter-domain topology disruption, end-to-end priority service request and virtual machine migration.

• The paper presents a distributed SDN control plane architecture that robustly manages multi-domain networks.
• It partitions control tasks into intra-domain and inter-domain modules to optimize resource management and reduce control traffic.
• Experimental evaluations demonstrate DISCO's ability to maintain service continuity under network disruptions and dynamic conditions.

DISCO: A Distributed Multi-Domain SDN Control Plane

The paper "DISCO: Distributed Multi-domain SDN Controllers" proposes a novel approach to design a distributed Software Defined Networking (SDN) control plane tailored for multi-domain environments. The work addresses the challenges of managing complex, large-scale networks that spread across various domains, including data centers, enterprise networks, and other sensitive applications. The authors, Kevin Phemius, Mathieu Bouet, and Jeremie Leguay, demonstrate how their architecture can enhance network robustness, scalability, and adaptability which are critical in environments where traditional SDN approaches fail to offer suitable solutions due to their inherent limitations such as single points of failure and lack of inter-domain communication efficiency.

Key Contributions

DISCO presents a distributed SDN control architecture where each controller oversees its own domain while maintaining inter-controller communications through a specialized channel. This differs from conventional distributed SDN solutions, which often rely on a consistent network-wide state that increases control traffic and is less adaptable to network heterogeneity. The paper leverages the Floodlight controller and the AMQP protocol to implement an architecture that emphasizes both intra-domain and inter-domain functionalities, providing agents responsible for monitoring, reachability, connectivity, and reservation.

Architectural Insights

The architecture splits the control tasks into intra-domain and inter-domain modules. Intra-domain modules manage local resources and provide functionalities such as path computation and flow management, taking advantage of an extended database that holds network states and statistics. Conversely, inter-domain modules enable communication between network domains, facilitating operations like end-to-end flow management and resource reservations.

Key elements of the control channel include:

1. Adaptive Information Exchange: To reduce unnecessary inter-controller traffic, DISCO features adaptive agents that adjust the frequency and volume of exchanged information based on network conditions, thereby alleviating congestion.
2. Service Reservation and Pre-emption: The system supports dynamic resource allocation based on service quality requirements, allowing high-priority flows to pre-empt resources initially allocated to lower-priority traffic.
3. Mobility Support: DISCO enhances network flexibility by supporting scenarios such as virtual machine migration, ensuring seamless communication continuity by dynamically updating path computations as host locations change.

Performance and Evaluation

Through experimental validations involving use cases like inter-domain topology disruption, the authors showcase DISCO's ability to maintain robust control under varied conditions. Notably, the system's adaptability is evidenced by its capability to manage control message overhead efficiently, shift routes in response to network failures, and uphold service-level agreements.

The paper provides quantitative analysis demonstrating that DISCO effectively mitigates controller failures and adapts to link disruptions without significant service interruption. Such empirical evidence suggests that DISCO can maintain low latency and packet loss rates even under dynamic network reconfigurations.

Implications and Future Work

The implications of this research are significant for the broader deployment and management of distributed SDN systems, particularly in environments characterized by diverse and geographically distributed network domains. DISCOS's architecture introduces a framework that others can extend and integrate with varied SDN protocols, fostering innovations in network resilience and autonomous operation.

The authors suggest investigating advanced control plane clustering techniques to improve scalability and efficiency. They also propose further enhancement of the interoperability between SDN and traditional IETF protocols, ensuring seamless integration of different networking paradigms without compromising performance or security.

This paper enriches the understanding of distributed SDN frameworks and sets the stage for continued exploration of sophisticated control mechanisms that address current limitations in network management at scale.