Das Chemnitzer StudentenNetz (CSN) setzt seit mehreren Jahren ein System zur automatischen Regelung der Bandbreite ("DynShaper") basierend auf den Grundlagen der Diplomarbeit von Jan Horbach "Dynamische Bandbreitenbeschränkung mit QoS" ein. Aufgrund der weiterhin bestehenden Notwendigkeit zum Einsatz dieses Systems, ist es nötig die DynShaper-Software zu überarbeiten, um ihre Implementierung an die Standards des CSN anzupassen und die unvollständige Dokumentation zu ergänzen. Hauptaugenmerk liegt dabei auf der Integration des Systems in die bestehende Softwarearchitektur des CSN und der Schaffung einer modularen Implementierung zur Evaluierung anderer Berechnungsverfahren.

After the complete liberalization of the airline industry during the 1990s the industry has faced a rapid growth in passenger numbers. This has mainly been caused by the emergence of so-called Low Cost Carrier (LCC) that offer a simplified product (i.e. point-to-point flights without any frills) at a lower cost than traditional Network Carriers. Furthermore LCC also introduced a less differentiated pricing structure (Restriction Free Pricing) which forced competing network carriers to reduce the degree of price discrimination which they were able to practice until then in order to defend their market shares. This has led to a decrease of average yields, which resulted in difficulties for (smaller) Network Carriers to cover their fixed costs, related to the operation of a hub & spoke network. In this environment network airlines are looking for new revenue sources as well as further sources of cost reduction. This development has amplified the consolidation trend of the airline industry and led to the emergence of several multi-hub networks (e.g. Lufthansa runs hub-operation in Frankfurt, Munich, Zurich and Vienna). One way to leverage the fact that multi-hub networks allow several routings for one origin-destination city pair would be the introduction of flexible tickets, where the actual routing of the passenger is not defined at the moment of purchase but only a certain time prior to departure. This allows airlines to raise the load factor on their network by increasing the degree of overbooking which they currently practice by pooling the risk that more passengers arrive than there is capacity among several flights. Furthermore these tickets might allow network carriers to compete in the low-cost-airline segment without having to further reduce the price level of their regular product (with specified routing). The present dissertation examined possible designs of such a ticket and their impact on the acceptance by passengers by means of a choice based conjoint study among 356 travelers. The findings suggest that while 77.5% of leisure travelers are willing to accept flexible time-range tickets in their relevant set, only 56% of business travelers are considering using this kind of ticket. More particular the results also showed that business travelers are not willing to compromise on travel duration and departure times, and are subsequently willing to pay a premium for specified tickets. A market share simulation showed that depending on the selected product layout flexible time-range tickets are able to gain up to 60% market share when offered at a discount of up to 33% relative to traditional tickets. When it comes to the actual layout, the largest lever to increase the acceptance is to exclude connection flights from the potential set of flights. The results contribute to the young research area on flexible products by assessing the disutility which is experienced by customers with regard to particular product characteristics of flexible products. Furthermore the results aim at providing airline managers with a comprehensive overview of the possibilities which flexible time-range tickets bring along when it comes to increasing the load factor and thereby the revenues in a multi-hub network. (author's abstract)

Universiẗat, Diss.--Göttingen, 2005. / Literaturverz. S. [297] - 312.

Universiẗat, Diss.--St. Gallen, 2005.

Zugl.: @Göttingen, Universiẗat, Diss., 2006.

Zugl.: Münster (Westfalen), Universiẗat, Diss., 2009.

In the pursuit for ever-increasing compute power, and with Moore's law slowly coming to an end, high-performance computing started to scale-out to larger systems. Alongside the increasing system size, the interconnection network is growing to accommodate and connect tens of thousands of compute nodes. These networks have a large influence on total cost, application performance, energy consumption, and overall system efficiency of the supercomputer. Unfortunately, state-of-the-art routing algorithms, which define the packet paths through the network, do not utilize this important resource efficiently. Topology-aware routing algorithms become increasingly inapplicable, due to irregular topologies, which either are irregular by design, or most often a result of hardware failures. Exchanging faulty network components potentially requires whole system downtime further increasing the cost of the failure. This management approach becomes more and more impractical due to the scale of today's networks and the accompanying steady decrease of the mean time between failures. Alternative methods of operating and maintaining these high-performance interconnects, both in terms of hardware- and software-management, are necessary to mitigate negative effects experienced by scientific applications executed on the supercomputer. However, existing topology-agnostic routing algorithms either suffer from poor load balancing or are not bounded in the number of virtual channels needed to resolve deadlocks in the routing tables. Using the fail-in-place strategy, a well-established method for storage systems to repair only critical component failures, is a feasible solution for current and future HPC interconnects as well as other large-scale installations such as data center networks. Although, an appropriate combination of topology and routing algorithm is required to minimize the throughput degradation for the entire system. This thesis contributes a network simulation toolchain to facilitate the process of finding a suitable combination, either during system design or while it is in operation. On top of this foundation, a key contribution is a novel scheduling-aware routing, which reduces fault-induced throughput degradation while improving overall network utilization. The scheduling-aware routing performs frequent property preserving routing updates to optimize the path balancing for simultaneously running batch jobs. The increased deployment of lossless interconnection networks, in conjunction with fail-in-place modes of operation and topology-agnostic, scheduling-aware routing algorithms, necessitates new solutions to solve the routing-deadlock problem. Therefore, this thesis further advances the state-of-the-art by introducing a novel concept of routing on the channel dependency graph, which allows the design of an universally applicable destination-based routing capable of optimizing the path balancing without exceeding a given number of virtual channels, which are a common hardware limitation. This disruptive innovation enables implicit deadlock-avoidance during path calculation, instead of solving both problems separately as all previous solutions.

Software Defined Networking (SDN) is revolutionizing the architecture and operation of computer networks and promises a more agile and cost-efficient network management. SDN centralizes the network control logic and separates the control plane from the data plane, thus enabling flexible management of networks. A network based on SDN consists of a data plane and a control plane. To assist management of devices and data flows, a network also has an independent monitoring plane. These coexisting network planes have various types of resources, such as bandwidth utilized to transmit monitoring data, energy spent to power data forwarding devices and computational resources to control a network. Unwise management, even abusive utilization of these resources lead to the degradation of the network performance and increase the Operating Expenditure (Opex) of the network owner. Conserving and protecting limited network resources is thus among the key requirements for efficient networking. However, the heterogeneity of the network hardware and network traffic workloads expands the configuration space of SDN, making it a challenging task to operate a network efficiently. Furthermore, the existing approaches usually lack the capability to automatically adapt network configurations to handle network dynamics and diverse optimization requirements. Addtionally, a centralized SDN controller has to run in a protected environment against certain attacks. This thesis builds upon the centralized management capability of SDN, and uses cross-layer network optimizations to perform joint traffic engineering, e.g., routing, hardware and software configurations. The overall goal is to overcome the management complexities in conserving and protecting resources in multiple functional planes in SDN when facing network heterogeneities and system dynamics. This thesis presents four contributions: (1) resource-efficient network monitoring, (2) resource-efficient data forwarding, (3) using self-adaptive algorithms to improve network resource efficiency, and (4) mitigating abusive usage of resources for network controlling. The first contribution of this thesis is a resource-efficient network monitoring solution. In this thesis, we consider one specific type of virtual network management function: flow packet inspection. This type of the network monitoring application requires to duplicate packets of target flows and send them to packet monitors for in-depth analysis. To avoid the competition for resources between the original data and duplicated data, the network operators can transmit the data flows through physically (e.g., different communication mediums) or virtually (e.g., distinguished network slices) separated channels having different resource consumption properties. We propose the REMO solution, namely Resource Efficient distributed Monitoring, to reduce the overall network resource consumption incurred by both types of data, via jointly considering the locations of the packet monitors, the selection of devices forking the data packets, and flow path scheduling strategies. In the second contribution of this thesis, we investigate the resource efficiency problem in hybrid, server-centric data center networks equipped with both traditional wired connections (e.g., InfiniBand or Ethernet) and advanced high-data-rate wireless links (e.g., directional 60GHz wireless technology). The configuration space of hybrid SDN equipped with both wired and wireless communication technologies is massively large due to the complexity brought by the device heterogeneity. To tackle this problem, we present the ECAS framework to reduce the power consumption and maintain the network performance. The approaches based on the optimization models and heuristic algorithms are considered as the traditional way to reduce the operation and facility resource consumption in SDN. These approaches are either difficult to directly solve or specific for a particular problem space. As the third contribution of this thesis, we investigates the approach of using Deep Reinforcement Learning (DRL) to improve the adaptivity of the management modules for network resource and data flow scheduling. The goal of the DRL agent in the SDN network is to reduce the power consumption of SDN networks without severely degrading the network performance. The fourth contribution of this thesis is a protection mechanism based upon flow rate limiting to mitigate abusive usage of the SDN control plane resource. Due to the centralized architecture of SDN and its handling mechanism for new data flows, the network controller can be the failure point due to the crafted cyber-attacks, especially the Control-Plane- Saturation (CPS) attack. We proposes an In-Network Flow mAnagement Scheme (INFAS) to effectively reduce the generation of malicious control packets depending on the parameters configured for the proposed mitigation algorithm. In summary, the contributions of this thesis address various unique challenges to construct resource-efficient and secure SDN. This is achieved by designing and implementing novel and intelligent models and algorithms to configure networks and perform network traffic engineering, in the protected centralized network controller.

Wertschöpfungsstrukturen in der Gesellschaft passen sich vor dem Hintergrund globaler Herausforderungen sehr flexibel und dynamisch an ständig neue Anforderungen an. Innovative Wertschöpfung kann dabei immer weniger von einzelnen Akteuren in Wissenschaft, Technik, Wirtschaft und Gesellschaft geleistet werden, sondern verschiedene interdisziplinäre Kompetenzen müssen gebündelt werden, die zielorientierte Vernetzung von Wissen und Ressourcen muss gestaltet werden. Unbestechlicher Maßstab für den Erfolg jeglicher wirtschaftlicher Unternehmung und damit auch von vernetzten Wertschöpfungsketten ist der Kunde mit seinen individuellen Bedürfnissen, das heißt letztlich die Akzeptanz und Absatzchancen von Produkten und Dienstleistungen am Markt.

The continuous transformation of the industrial society into a service and knowledge society is accompanied by profound change of demand: Customer requests will increasingly focus on individual products, shorter delivery times and appropriate prices. To encounter these challenges under the conditions of dynamic global markets and inter-regional competition, the CRC 457 “Non-Hierarchical Regional Production-Networks” at the Chemnitz University of Technology focuses on SME and investigates ways to implement customer-oriented, temporary networking of smallest, autonomous units (“competence cells”) in the region of South-West-Saxony. As a partial result of this research we are able to present a model of competence cell based networking distinguishing between three levels of networking, each with its own logic of reproduction (cp. Figure 1): 1) According to a most general concept, networks can be seen as given in any social context. Regarding a specific region, one has to assume that there is a broad array of infra-structural, mental and communicative relations; that as a whole we call the Regional Network. This level can be seen as basic level of competence cell based networking, specific elements (competence cells) of this network can be seen as necessary resources of the following level. 2) The second level of networking we call Competence Network. The first continuous problem of this network is to select competence cells according to relevant parameters that may be defined by customer orders or, in view of new market entry, be generated by the Competence Network itself. The second problem is to arrange these cells along a product specific value chain and, by this means, to create temporary production networks. Concerning both of the problems, the third is to find a non-hierarchical mode of negotiation. 3) On the third level temporary alliances of basically autonomous competence cells can be observed. After executing their order, these Production Networks dissolve and the constituent cells reenter the second or even the first level of networking. Generalising our observation we now assume that every phenomena labelled as regional or industrial network or district should be investigated in view of (these) different levels of networking, especially if we are interested in dealing more efficiently with the problems of network management. Hence, we suggest to first answer the question “Which (of these) networks are we to manage?” before asking how. Concerning the latter, a synopsis on the discourse on managing networks shows three major approaches: 1) Networks are assumed to be uncontrollable (cp. e.g. Castells 2001), as they are self-organised systems selecting external control impulses only by their own criterias of relevance, so we have either to adapt the logic of the target system (which implies self-adaption) or to take the risk of being totally ignored. With this in mind we can’t speak of control in terms of causal logic. 2) Networks are assumed to be cultivable: According to Wenger and Snyder (2000) for example networks are uncontrollable, but - like a gardener - we are able to set adequate general conditions for their “growth” and continuity. 3) Networks are assumed to be controllable: For authors like Sydow (2000) networks are to be seen as hybrid forms of coordination recombining aspects of market and hierarchy. Thus, recombining the classical means of control will finally lead to the ability to manage networks. It will be to show that each of these approaches focuses on specific aspects of networking, thus, each of them is creating a self-contained concept of networks: Castells refers to networks as a basal nexus of (global) interconnectedness that, like the Regional Network in our case, obviously can’t be managed. Wenger and Snyder are interested in network-organisations that (like Competence Networks) also can’t be controlled but can be implemented by and embedded in other social systems (cp. Grabher 1993). Sydow finally observes networks of more or less autonomous organisations corresponding to our Production Networks, which are in deed more or less controllable. Not least because of their different theoretical assumptions, all these concepts seemed to be not only incommensurable but incompatible so far. Nonetheless, during our work in CRC 457 we were to manage all these concepts while approaching to one single objective of research. Therefore, we had to ask, how many perspectives on network are we to manage? The answer is: at least three. Integrating these, we generated a holistic life-cycle model of network, which we like to present as starting point for further research in the field network management.