Industrial networks require real-time guarantees for the flows they carry. That is, flows have hard end-to-end delay requirements that have to be deterministically guaranteed. While proprietary extensions of Ethernet have provided solutions, these often require expensive forwarding devices. The rise of Software-Defined Networking (SDN) opens the door to the design of centralized traffic engineering frameworks for providing such real-time guarantees. As part of such a framework, a network model is needed for the computation of worst-case delays and for access control. In this article, we propose two network models based on network calculus theory for providing deterministic services (DetServ). While our first model, the multi-hop model (MHM), assigns a rate and a buffer budget to each queue in the network, our second model, the threshold-based model (TBM), simply fixes a maximum delay for each queue. Via a packet-level simulation, we confirm that the delay bounds guaranteed by both models are never exceeded and that no packet loss occurs. We further show that the TBM provides more flexibility with respect to the characteristics of the flows to be embedded and that it has the potential of accepting more flows in a given network. Finally, we show that the runtime cost for this increase in flexibility stays reasonable for online request processing in industrial scenarios.
State synchronisation in clustered Software Defined Networking controller deployments ensures that all instances of the controller have the same state information in order to provide redundancy. Current implementations of controllers use a strong consistency model, where configuration changes must be synchronised across a number of instances before they are applied on the network infrastructure. For large deployments, this blocking process increases the delay of state synchronisation across cluster members and consequently has a detrimental effect on network operations that require rapid response, such as fast failover and Quality of Service applications. In this paper, we introduce an adaptive consistency model for SDN Controllers that employs concepts of eventual consistency models along with a novel 'cost-based' approach where strict synchronisation is employed for critical operations that affect a large portion of the network resources while less critical changes are periodically propagated across cluster nodes. We use simulation to evaluate our model and demonstrate the potential gains in performance.
Abstract-With the emergence of Software-Defined Networking (SDN) and Network Function Virtualization (NFV), the problem of centralized routing through intermediate specified nodes (for which several candidates can be defined) has become an important issue. Indeed, an SDN controller routing flows through Service Function Chains (SFCs) has to efficiently solve this problem to achieve the online provisioning of routing requests and the online placement of Virtual Network Functions (VNFs). In this paper, we propose two algorithms for solving this problem. First, we propose LARAC for specified nodes (LARAC-SN), a fast and close to optimal algorithm for finding the constrained shortest path (CSP) visiting an ordered set of specified nodes. Second, we propose Mole in the Hole (MITH), a graph transformation algorithm which can force any state-of-the-art routing algorithm to visit an ordered set of specified nodes. While LARAC-SN is bounded to the specific CSP problem and can only handle one candidate per specified node, MITH can be used for any routing problem and can deal with several candidates per specified node. Through evaluations, we show that LARAC-SN is fast and close to optimal (its optimality gap stays lower than 1.62% in average) and that MITH has the potential of reaching optimality for any problem, but at the cost of a higher runtime.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.