Zoran Despotovic

A novel approach to virtual networks embedding for SDN management and orchestration

The development of methodologies to manage and orchestrate virtualised resources and network functions is a fundamental enabler for optimally utilising physical ICT infrastructures. Algorithms for optimal location (embedding) of network functions, IT and CT resources, services and corresponding states, especially at the network edge, will enable new business models and provide a key competitive advantage to network administrators. This paper introduces a novel Mixed Integer Programming (MIP) formulation for a coordinated node and link mapping onto the underlying network infrastructure. Extensive simulation results show that the proposed algorithm outperforms prior art formulations: two digit gains were attained in terms of resources utilisation, embedding, revenues and, especially convergence time. The proposed methodology is applicable to a number of relevant use cases, as constraints and objective functions can be flexibly defined by network operators.

Virtual Links Mapping in Future SDN-Enabled Networks

Software defined networking (SDN) has emerged as an efficient network technology for lowering operating costs through simplified hardware, software and management. Specific research focus has been placed to achieve a successful carrier grade network with SDN, in terms of scalability, reliability, QoS and service management. In the literature, very little material is currently available on traffic engineering (TE) using this technology. This paper presents a novel mixed integer linear programming (MILP) formulation for a centralised controller to calculate optimal end-to-end virtual paths over the underlying network infrastructure, considering multiple requests simultaneously. Extensive simulation results, over a wide range of underlying network topologies and input parameters, demonstrate that the proposed algorithm outperforms traditional shortest path first (SPF) approaches. In some cases, up to 30 % more virtual connections were satisfactorily mapped onto the same substrate, independent of the number of physical nodes.

Understanding processing latency of SDN based mobility management in mobile core networks

Current solutions to evolve mobility management in Mobile Core Networks (MCN) based on SDN have showed the gain in flexibility and how to reduce control signaling. However, these works neglect the problems of describing the design choices of SDN Mobility Management Applications (MMA) and identifying where and what are the critical processing latency contributors for such design. Our paper addresses these problems. We study the internal mechanisms and interactions of MMA and controller, to determine the contributors to the overall processing latency. We implemented two MMA solutions (based on reactive and proactive designs) as modules of Floodlight, and we run experiments using Mininet and OpenFlow. The proactive MMA design can guarantee that the overall processing latency in the 95th percentile can be kept near the median value, given available CPU capacity at the SDN controller. One key lesson learned with our study is that only optimizing control signaling of MMA is not enough to provide better overall processing latency.

Modeling Reliability Requirements in Coordinated Node and Link Mapping

High performance systems require high levels of reliability. Many functions involved in telecommunication and IT networks have reliability requirements that can only be achieved by introducing redundant resources. In the telecom sector, recently, there has been a significant effort on moving carrier grade systems and functions to virtualized network infrastructure. The management and coordination of those virtualized systems to achieve an optimal mapping (or embedding) to the physical resources that host them is known as virtual resource orchestration. In our prior work we introduced a novel model, based on Mixed Integer Programming (MIP) problem formulation, as one significant way of achieving this optimality in embedding. This paper extends the model to include reliability requirements, improving prior art techniques as well as implementing a novel approach, denoted as reliability assurance. The confidence of the target reliability of the embedded virtual graphs can be traded with the efficiency of the substrate utilization. Extensive simulation results show that our model provides embedding rates and infrastructure utilization comparable with prior art while fulfilling high reliability requirements.

Dispatching PACKET_INs to the right SDN control application via context interpretation in Mobile Core Networks

Telco operators started to apply the SDN technologies also in the design of Mobile Core Networks (MCNs). In this change towards SDNized Mobile Core Network, it is crucial to understand how conventional interfaces among different mobile network entities should evolve. The issues derived from this change have not been tackled by current research approaches. This paper presents the first initiative to close this research gap. We tackle the key problem of how to identify which mobile SDN applications (APPs) should be invoked once a PACKET_IN (the OpenFlow message that transport information from the data plane to the control plane) is received at the control level. We propose data structures, a model, and detailed examples of three important PACKET_IN context interpretation for MCNs. Initial experiments, based on Floodlight controller and Mininet emulation environment were carried out. The results indicate that it is feasible to use our proposed approach to dispatch PACKET_INs to the right SDN APP. The delay introduced due to invocation of such mechanism to interpret the context of the PACKET_IN and activate the appropriate mobile SDN APPs is only in the order of microseconds. Our proposal can be used to simplify current Mobile Core Network interface design by exploiting the SDN mechanisms. We believe, this work helps to pave the way towards fully SDNized Mobile Core Networks.

Identifying latency factors in SDN-based Mobile Core Networks

Software Defined Networking (SDN) is considered one of the major driving factors to bring radical changes on Mobile Core Networks (MCN) design. There are many proposals on how to advocate SDN methodology, however, most of them stop at “vision” level without providing details about key performance contributors. In this paper, we tackle this gap and present a study of latency in SDN based MCN. We identify the major factors and elements in an SDN MCN that contribute to the overall latency. Two types of latency are considered: the processing delay inside the SDN control plane and the transmission delay of SDN control messages between controller and the managed switches. We use a realistic system setup with implementations of SDN controller and SDN application, as well as in-band transfer of control messages among switches and controller. We compare the latency obtained from SDN based MCN with the Evolved Packet Core (EPC). The observed overall latency in our experiments for SDN based MCN is within the EPC requirements. The direct comparison of the SDN versus EPC based MCN in the proposed scenarios show the first can reduce the overall latency. We also identified that the calls to controller APIs can be the major key contributors to the overall latency in SDN MCN, but the act of transmitting more control messages from the controller to the switches to setup longer flowpaths is not a key contributor for the overall latency.

On optimal hierarchical SDN

To address scalability concerns, hierarchical control plane organization has been proposed for SDN. However, an open question is how such a hierarchy should look like. In this paper, we model the impact of hierarchies on control plane performance and derive an expression for an optimal hierarchical organization for a given network scale. We then show that using a 4-layer SDN is sufficient for practical network scales, suggesting feasibility and relevance of this approach. Finally, we illustrate the elasticity of hierarchical SDN, which enables a more flexible cost evolution of the SDN control plane.

VNetMapper: A fast and scalable approach to virtual networks embedding

Virtual network embedding is considered an important problem to solve in order to make infrastructure virtualization economically reasonable. The most efficient algorithms proposed so far define link and node mappings as optimal solutions of Integer Programming (IP) problems. They exhibit reasonably good performance only for small problem instance sizes, including few tens of nodes and links per physical substrate, few nodes and links per virtual request and a dozen of virtual requests to handle in parallel. However, we find these instances too small to be of any practical use. To address this scalability issue, we propose VNetMapper, an algorithm to solve the virtual network embedding problem based on an integer program formulation with appropriately selected objective function, variables and the set of constraints tuned to give an optimal performance. We show through simulations that VNetMapper can quickly, within seconds, solve large problem instances, involving physical substrates with hundreds of nodes and thousands of links and batches of hundreds of virtual network requests. By identifying exact properties that make VNetMapper so fast and scalable, we present guidelines for designing scalable integer programs.

SDN on ACIDs

Software-defined networks (SDN) do not guarantee coherent network operations, when uncoordinated SDN applications concurrently update the network forwarding state. As this problem has not so far received considerable attention, in this paper, we introduce transactional network-wide update support in SDN. We notably design and implement a new SDN service complementing state of the art proposals to achieve atomicity, isolation and durability of network updates in any typical SDN setup. The experiments with our implementation of the ACID service in FloodLight and OpenVSwitch demonstrate the practical feasibility of our proposal and good scalability to different network loads and sizes.

Autonomic load balancing in the future disintegrated and virtualized networks

With the emergence of the concepts like virtualization and cloud computing, the solutions based on disintegration and virtualization of network resources have also gained immense attention in the research community. These solutions focus on a dynamic and fine-grained network management. In this paper, we investigate the autonomic network management (specifically in congested resource scenarios) in fully virtualized and disintegrated network paradigm. We model the problem of congestion avoidance using cooperative game-theoretic approach. We also extensively implement the validation framework and deploy the proposed algorithm on the developed validation platform. We observe that the proposed algorithm autonomically balances the load and reflects gain in terms of resource utilization.