Matthias Falkner

An overview of pricing concepts for broadband IP networks

In this article we provide an overview of pricing concepts for broadband multiservice networks. We review the notions of flat pricing, priority pricing, Paris-Metro pricing, smart-market pricing, responsive pricing, expected capacity pricing, edge pricing, and effective bandwidth pricing. We use numerous evaluation criteria, including network, economic, and social efficiency, as well as their suitability in using pricing as a means for congestion control. Some of the schemes are based on best-effort networks, and are thus unable to provide the user with quality of service (QoS) guarantees. Others build on networks with connection admission control functions and are thus able to provide individual QoS guarantees. We particularly investigate the relevant time frame over which pricing schemes are assumed to operate. The majority of the schemes work on short time frames (on the order of minutes), which makes them applicable to use pricing as an additional means for controlling congestion. We also consider technical aspects such as compliance with existing networking technologies or computational overheads associated with billing and accounting.

Fast simulation of networks of queues with effective and decoupling bandwidths

A significant difficulty when using Monte Carlo simulation for the performance analysis of communication networks is the long runtime required to obtain accurate statistical estimates. Under the proper conditions, importance sampling (IS) is a technique that can speed up simulations involving rare events in network (queuing) systems. Large speed-up factors in simulation runtime can be obtained with IS if the modification or bias of the underlying probability measures of certain random processes is carefully chosen. Fast simulation methods based on large deviation theory (LTD) have been successfully applied in many cases. In this paper, we set up an IS-based simulation of various elementary network topologies. These configurations are frequenly encountered in broadband ATM-based network components such as switches and multiplexers. Our objective in this study is to obtain the optimal or near-optimal biasing parameter values of the arrival processes for the importance sampling simulation. For this purpose we appropriately apply a technique presented by Chang et al. for certain portions of the networks (intree) while we develop a new algorithm, inspired by the work of De Veciana et al. on decoupling bandwidths, for the non-intree portion of the network.

Minimum cost traffic shaping: a user's perspective on connection admission control

We propose a minimum cost method for traffic shaping in the context of QoS-based networks. Given the users desired QoS and the networks resource availability, our procedure determines the least-cost parameters for a traffic shaper which still guarantees access to the network whilst satisfying the QoS constraints. We illustrate our scheme using on-off sources and formulate the QoS constraints by effective bandwidths.

Optimal virtualized network function allocation for an SDN enabled cloud

Enterprise and Service Provider networks are increasingly making use of Virtualized Network Functions (VNFs) to reap the benefits of reduced Capital expenditures (CAPEX) and Operating expenses (OPEX). Total cost of ownership calculations however are typically a function of the attainable network performance, which in a virtualized system is highly dependent on the overall system architecture and the resource allocation policy. This paper examines how to improve the overall performance of deploying VNFs in a virtualized environment, proposing an appropriate architectural framework to leverage the capabilities of other technologies such as Software Defined Networking (SDN) and Cloud Computing which can be highly complemented with the Network Function Virtualization (NFV) paradigm. We focus on the standardization of the three technologies under a common architecture and we incorporate resource allocation as a main functional architectural block. The problem studied is referred as the VNF placement problem, and tries to efficiently, intelligently, and dynamically allocate a series of VNFs comprising a service chain in a cloud environment according to the current status of the cloud resources. We propose and evaluate a set of algorithms towards satisfying the requirement of enterprise and service providers to grant high performance VNFs for the end users, while allowing the cloud providers to profitably utilize the available resources. Graphical abstractDisplay Omitted HighlightsDynamic traffic steering and allocation of a service chain in an SDN enabled Cloud.Service Chain Deployment Models.Standardization of NFV, SDN, and Cloud Computing.Optimal VNF allocation based in Mixed Integer Programming.Heuristic Algorithms to approximate optimal solution.

Performance analysis of virtualized network functions on virtualized systems architectures

Network Function Virtualization (NFV) is an emerging network architecture that employs the concept of virtualization and allows the consolidation of many network services on top of an industry standard off-the-shelf server. This decoupling of network functions and services from dedicated and expensive hardware appliances has led the Enterprise and Service Providers to increasingly make use of Virtualized Network Functions (VNFs) to reap the benefits of reduced capital and operational expenses. Total cost of ownership calculations however are typically a function of the attainable network throughput and performance, which in a virtualized system is highly dependent on the overall system architecture. The number of VNFs running on the server, their I/O demands, the performance characterization of the underlying hypervisor scheduler, or the packet path from physical interfaces into the VNFs are examples of how the system architecture can influence overall performance and throughput. This article provides the challenges of deploying VNFs on a virtualized system architecture and analyzes the impact of the architecture on the overall VNF performance under both single-VNF and multi-VNF configurations.

Optimal cost traffic shaping with self-similar input sources

We address the problem of determining the parameters of a traffic shaper modeled as a single server queue with finite buffer size and deterministic service rate. In our scheme, the user wishes to gain access to the network with cell loss probability guarantees when the input traffic is self-similar. We assume that the mean rate, the variance and the degree of self-similarity are known. The network is able to provide a connection if the user voluntarily shapes the traffic and if sufficient resources are available to accommodate the shaped traffic stream. We formulate a minimization problem to determine the optimal parameters for the traffic shaper and use techniques of non-linear programming to obtain a solution.

Minimum cost traffic shaping

We propose a minimum-cost method for traffic shaping in the context of quality of service (QoS)-based networks. Given the users desired QoS and the networks resource availability, our method determines the least cost parameters for a shaper while guaranteeing access to the network and satisfying the QoS requirements.

Multi-VNF performance characterization for virtualized network functions

Network Function Virtualization promises to reduce the overall operational and capital expenses experienced by the network operators. Running multiple network functions on top of a standard x86 server instead of dedicated appliances can increase the utilization of the underlying hardware and reduce the maintenance and management costs. However, total cost of ownership calculations are typically a function of the attainable network throughput, which in a virtualized system is highly dependent on the overall system architecture - in particular the input/ output (I/O) path. In this paper, we investigate the attainable performance of an x86 host running multiple Virtualized Network Functions (VNFs) under different I/O architectures: OVS, SRIOV and FD.io VPP. We show that the system throughput in a multi-VNF environment differs significantly from deployments where only a single VNF is running on a server, while different I/O architectures can achieve different levels of performance.

Resource Management and Orchestration for a Dynamic Service Chain Steering Model

Network Function Virtualization along with Network Service Chaining envision a reduction in the respective cost that end users, service providers, and network operators are experiencing, while providing complete and high quality services. However, the vast range of available services and the service on-demand model, creates dynamic traffic conditions that necessitates a flexible and automatic network platform to redirect traffic according to network conditions. In this paper, we study the problem of deploying service chains, consisting of a number of virtualized network functions (VNFs), in a SDN enabled data center network, where a random number of users are associated with each service chain. To this end, appropriate resource management algorithms are introduced for the placement of VNFs satisfying server affinity and latency constraints. The interconnection of the VNFs is facilitated by an SDN controller, which periodically recalculates the routing paths to adjust to the dynamic traffic conditions.

Dynamic traffic steering of multi-tenant virtualized network functions in SDN enabled data centers

Network Functions Visualization (NFV) builds upon virtualization technology and allows the decoupling of network services from dedicated and complex hardware-based appliances. This decoupling allows network providers to install these functions in commercial off-the-shelf hardware reducing the overall capital and operational expenses. In order to provide complete and high quality services to the end users, a series of these functions may need to be traversed creating a service chain model. In this paper, we study the problem of allocating a dynamic service chain, where a random number of users are associated with each service chain, in a data center network. The allocation consists of the placement of VNFs in appropriate servers and the interconnection between these servers. This interconnection is facilitated by an SDN controller, which proves to be complementary with the NFV paradigm. Towards these aims, we propose two algorithms for the placement of the VNFs, while the SDN controller periodically recalculates the routing paths to adjust to the dynamic traffic conditions.