arXiv subjects
Publications and source records attributed to Wolfgang Kellerer.
Reproducibility is one of the key characteristics of good science, but hard to achieve for experimental disciplines like Internet measurements and networked systems. This guide provides advice to researchers, particularly those new to the field, on designing experiments so that their work is more likely to be reproducible and to serve as a foundation for follow-on work by others.
The advent of Machine-to-Machine communication has sparked a new wave of interest to random access protocols, especially in application to LTE Random Access (RA). By analogy with classical slotted ALOHA, state-of-the-art models LTE RA as a multi-channel slotted ALOHA. In this letter, we direct the attention to the resource consumption of RA. We show that the consumption is a random variable dependent on the contention parameters. We consider two approaches to include the consumption into RA optimization: by defining resource efficiency and by the means of a bi-objective optimization, where resource consumption and throughput are the competing objectives. We then develop the algorithm to obtain Pareto-optimal RA configuration under resource constraint. We show that the algorithm achieves lower burst resolution delay and higher throughput than the state-of-the-art.
Mobile video consumption is increasing and sophisticated video quality adaptation strategies are required to deal with mobile throughput fluctuations. These adaptation strategies have to keep the switching frequency low, the average quality high and prevent stalling occurrences to ensure customer satisfaction. This paper proposes a novel methodology for the design of machine learning-based adaptation logics named HASBRAIN. Furthermore, the performance of a trained neural network against two algorithms from the literature is evaluated. We first use a modified existing optimization formulation to calculate optimal adaptation paths with a minimum number of quality switches for a wide range of videos and for challenging mobile throughput patterns. Afterwards we use the resulting optimal adaptation paths to train and compare different machine learning models. The evaluation shows that an artificial neural network-based model can reach a high average quality with a low number of switches in the mobile scenario. The proposed methodology is general enough to be extended for further designs of machine learning-based algorithms and the provided model can be deployed in on-demand streaming scenarios or be further refined using reward-based mechanisms such as reinforcement learning. All tools, models and datasets created during the work are provided as open-source software.
The Network Function Virtualization (NFV) is very promising for efficient provisioning of network services and is attracting a lot of attention. NFV can be implemented in commercial off-the-shelf servers or Physical Machines (PMs), and many network services can be offered as a sequence of Virtual Network Functions (VNFs), known as VNF chains. Furthermore, many existing network devices (e.g., switches) and collocated PMs are underutilized or over-provisioned, resulting in low power-efficiency. In order to achieve more energy efficient systems, this work aims at designing the placement of VNFs such that the total power consumption in network nodes and PMs is minimized, while meeting the delay and capacity requirements of the foreseen demands. Based on existing switch and PM power models, we propose a Integer Linear Programming (ILP) formulation to find the optimal solution. We also propose a heuristic based on the concept of Blocking Islands (BI), and a baseline heuristic based on the Betweenness Centrality (BC) property of the graph. Both heuristics and the ILP solutions have been compared in terms of total power consumption, delay, demands acceptance rate, and computation time. Our simulation results suggest that BI-based heuristic is superior compared with the BC-based heuristic, and very close to the optimal solution obtained from the ILP in terms of total power consumption and demands acceptance rate. Compared to the ILP, the proposed BI-based heuristic is significantly faster and results in 22% lower end-to-end delay, with a penalty of consuming 6% more power in average.
The routing algorithms used by current operators aim at coping with the demanded QoS requirements while optimizing the use of their network resources. These algorithms rely on the optimal substructure property (OSP), which states that an optimal path contains other optimal paths within it. However, we show that QoS metrics such as queuing delay and buffer consumption do not satisfy this property, which implies that the used algorithms lose their optimality and/or completeness. This negatively impacts the operator economy by causing a waste of network resources and/or violating Service Level Agreements (SLAs). In this paper, we propose a new so-called Mn taxonomy defining new metric classes. An Mn metric corresponds to a metric which requires the knowledge of the n previously traversed edges to compute its value at a given edge. Based on this taxonomy, we present three solutions for solving routing problems with the newly defined classes of metrics. We show that state-of-the-art algorithms based on the OSP indeed lose their original optimality and/or completeness properties while our proposed solutions do not, at the price of an increased computation time.
Random Access Channel (RACH) has been identified as one of the major bottlenecks for accommodating massive number of machine-to-machine (M2M) users in LTE networks, especially for the case of burst arrival of connection requests. As a consequence, the burst resolution problem has sparked a large number of works in the area, analyzing and optimizing the average performance of RACH. However, the understanding of what are the probabilistic performance limits of RACH is still missing. To address this limitation, in the paper, we investigate the reliability of RACH with access class barring (ACB). We model RACH as a queuing system, and apply stochastic network calculus to derive probabilistic performance bounds for burst resolution time, i.e., the worst case time it takes to connect a burst of M2M devices to the base station. We illustrate the accuracy of the proposed methodology and its potential applications in performance assessment and system dimensioning.
Datacenter operators and internet content providers require optical network programmability to efficiently interconnect distributed datacenters. This paper describes the requirements, applications and use cases for optical network programmability. Based on application scenarios, open northbound APIs with different levels of control and abstraction to address different network operator requirements which are defined. For their illustration, use cases for optical network programmability show current research directions.
Growing traffic demands and increasing security awareness are driving the need for secure services. Current solutions require manual configuration and deployment based on the customer's requirements. In this work, we present an architecture for an automatic intent-based provisioning of a secure service in a multilayer - IP, Ethernet, and optical - network while choosing the appropriate encryption layer using an open-source software-defined networking (SDN) orchestrator. The approach is experimentally evaluated in a testbed with commercial equipment. Results indicate that the processing impact of secure channel creation on a controller is negligible. As the time for setting up services over WDM varies between technologies, it needs to be taken into account in the decision-making process.
Software-defined networking is finding its way into optical networks. Here, it promises a simplification and unification of network management for optical networks allowing automation of operational tasks despite the highly diverse and vendor-specific commercial systems and the complexity and analog nature of optical transmission. A fundamental component for software-defined optical networking are common abstractions and interfaces. Currently, a number of models for optical networks are available. They all claim to provide open and vendor agnostic management of optical equipment. In this work, we survey and compare the most important models and propose an intent interface for creating virtual topologies that is integrated in the existing model ecosystem.
Accommodating Machine-to-Machine applications and their requirements is one of the challenges on the way from LTE towards 5G networks. The envisioned high density of devices, alongside with their sporadic and synchronized transmission patterns, might create signaling storms and overload in the current LTE network. Here, we address the notorious random access (RA) challenge, namely, scalability of the radio link connection establishment protocol in LTE networks. We revisit the binary countdown technique for contention resolution (BCCR), and apply it to the LTE RA procedure. We analytically investigate the performance gains and trade-offs of applying BCCR in LTE. We further simulatively compare BCCR RA with the state-of-the-art RA techniques, and demonstrate its advantages in terms of delay and throughput.
Network hypervisors provide the network virtualization layer for Software Defined Networking (SDN). They enable virtual network (VN) tenants to bring their SDN controllers to program their logical networks individually according to their demands. In order to make use of the high flexibility of virtual SDN networks and to provide high performance, the deployment of the virtualization layer needs to adapt to changing VN demands. This paper initializes the study of the optimization of dynamic SDN network virtualization layers. Based on the definition of reconfiguration events, we formalized mixed integer programs to analyze the multi-objective problem of adapting virtualization layers. Our initial simulation results demonstrate Pareto frontiers of conflicting objectives, namely control plane latency and hypervisor and control path reconfigurations.
Network Functions Virtualization (NFV) is a promising network architecture where network functions are virtualized and decoupled from proprietary hardware. In modern datacenters, user network traffic requires a set of Virtual Network Functions (VNFs) as a service chain to process traffic demands. Traffic fluctuations in Large-scale DataCenters (LDCs) could result in overload and underload phenomena in service chains. In this paper, we propose a distributed approach based on Alternating Direction Method of Multipliers (ADMM) to jointly load balance the traffic and horizontally scale up and down VNFs in LDCs with minimum deployment and forwarding costs. Initially we formulate the targeted optimization problem as a Mixed Integer Linear Programming (MILP) model, which is NP-complete. Secondly, we relax it into two Linear Programming (LP) models to cope with over and underloaded service chains. In the case of small or medium size datacenters, LP models could be run in a central fashion with a low time complexity. However, in LDCs, increasing the number of LP variables results in additional time consumption in the central algorithm. To mitigate this, our study proposes a distributed approach based on ADMM. The effectiveness of the proposed mechanism is validated in different scenarios.
Contention tree algorithm is initially invented as a solution to improve the stable throughput problem of Slotted ALOHA in multiple access schemes. Even though the throughput is stabilized in tree algorithms, the delay of requests may grow to infinity with respect to the arrival rate of the system. Delay depends heavily on the exploration of the tree structure, i.e., breadth search, or depth search. Breadth search is necessary for faster exploration of tree. The analytical probability distribution of delay, which is available mostly for depth search, is not generalizable to all breadth search. In this paper we fill this gap through though arbitrary grouping of branches and including this in the delay analysis. This enables obtaining the delay analysis of any contention tree algorithm that runs a breadth first search exploration. We show through simulations that the analysis is in agreement with the realizations.
Ideally, by enabling multi-tenancy, network virtualization allows to improve resource utilization, while providing performance isolation: although the underlying resources are shared, the virtual network appears as a dedicated network to the tenant. However, providing such an illusion is challenging in practice, and over the last years, many expedient approaches have been proposed to provide performance isolation in virtual networks, by enforcing bandwidth reservations. We in this paper study another source for overheads and unpredictable performance in virtual networks: the hypervisor. The hypervisor is a critical component in multi-tenant environments, but its overhead and influence on performance are hardly understood today. In particular, we focus on OpenFlow-based virtualized Software Defined Networks (vSDNs). Network virtualization is considered a killer application for SDNs: a vSDN allows each tenant to flexibly manage its network from a logically centralized perspective, via a simple API. For the purpose of our study, we developed a new benchmarking tool for OpenFlow control and data planes, enabling high and consistent OpenFlow message rates. Using our tool, we identify and measure controllable and uncontrollable effects on performance and overhead, including the hypervisor technology, the number of tenants as well as the tenant type, as well as the type of OpenFlow messages.
This paper addresses network anomography, that is, the problem of inferring network-level anomalies from indirect link measurements. This problem is cast as a low-rank subspace tracking problem for normal flows under incomplete observations, and an outlier detection problem for abnormal flows. Since traffic data is large-scale time-structured data accompanied with noise and outliers under partial observations, an efficient modeling method is essential. To this end, this paper proposes an online subspace tracking of a Hankelized time-structured traffic tensor for normal flows based on the Candecomp/PARAFAC decomposition exploiting the recursive least squares (RLS) algorithm. We estimate abnormal flows as outlier sparse flows via sparsity maximization in the underlying under-constrained linear-inverse problem. A major advantage is that our algorithm estimates normal flows by low-dimensional matrices with time-directional features as well as the spatial correlation of multiple links without using the past observed measurements and the past model parameters. Extensive numerical evaluations show that the proposed algorithm achieves faster convergence per iteration of model approximation, and better volume anomaly detection performance compared to state-of-the-art algorithms.
The emerging Software Defined Networking (SDN) paradigm separates the data plane from the control plane and centralizes network control in an SDN controller. Applications interact with controllers to implement network services, such as network transport with Quality of Service (QoS). SDN facilitates the virtualization of network functions so that multiple virtual networks can operate over a given installed physical network infrastructure. Due to the specific characteristics of optical (photonic) communication components and the high optical transmission capacities, SDN based optical networking poses particular challenges, but holds also great potential. In this article, we comprehensively survey studies that examine the SDN paradigm in optical networks; in brief, we survey the area of Software Defined Optical Networks (SDONs). We mainly organize the SDON studies into studies focused on the infrastructure layer, the control layer, and the application layer. Moreover, we cover SDON studies focused on network virtualization, as well as SDON studies focused on the orchestration of multilayer and multidomain networking. Based on the survey, we identify open challenges for SDONs and outline future directions.
Flexibility is often claimed as a competitive advantage when proposing new network designs. However, most proposals provide only qualitative arguments for their improved support of flexibility. Quantitative arguments vary a lot among different proposals. A general understanding for flexibility is not yet clearly defined, leaving it to the reader to draw the right conclusions based on background information. The term flexibility is commonly defined as the ability to adapt to changes. For networks, flexibility would refer to the ability to adapt the available network resources, such as flows or topology, to changes of design requirements, e.g., shorter latency budgets or different traffic distributions. Recent concepts such as Software Defined Networking, Network Virtualization and Network Function Virtualization have emerged claiming to provide more flexibility in networks. Nevertheless, a deeper understanding of what flexibility means and how it could be quantified to compare different network designs remains open. In this paper, we ask whether flexibility can be a new measure for network design space analysis. As it is quite challenging to formulate a flexibility measure that covers all network aspects, we propose an initial set of flexibility aspects to start grounding guidelines. Our initial selection is backed up by an analysis of Software Defined Networking, Network Virtualization and Network Function Virtualization for their support of the selected flexibility aspects. Our research methodology is based on a systematic approach that leads to network design guidelines with respect to flexibility.
Software defined networking (SDN) has emerged as a promising paradigm for making the control of communication networks flexible. SDN separates the data packet forwarding plane, i.e., the data plane, from the control plane and employs a central controller. Network virtualization allows the flexible sharing of physical networking resources by multiple users (tenants). Each tenant runs its own applications over its virtual network, i.e., its slice of the actual physical network. The virtualization of SDN networks promises to allow networks to leverage the combined benefits of SDN networking and network virtualization and has therefore attracted significant research attention in recent years. A critical component for virtualizing SDN networks is an SDN hypervisor that abstracts the underlying physical SDN network into multiple logically isolated virtual SDN networks (vSDNs), each with its own controller. We comprehensively survey hypervisors for SDN networks in this article. We categorize the SDN hypervisors according to their architecture into centralized and distributed hypervisors. We furthermore sub-classify the hypervisors according to their execution platform into hypervisors running exclusively on general-purpose compute platforms, or on a combination of general-purpose compute platforms with general- or special-purpose network elements. We exhaustively compare the network attribute abstraction and isolation features of the existing SDN hypervisors. As part of the future research agenda, we outline the development of a performance evaluation framework for SDN hypervisors.