Bibliothek

Beschreibung: In dieser Arbeit werden effiziente Co-Reservierungs-Algorithmen vorgestellt, mit denen Rechenressourcen mehrerer Cluster im Grid reserviert werden können. Es werden fünf Algorithmen vorgestellt, die flexible Co-Reservierungsanfragen ermöglichen. In einer Co-Reservierungsanfrage müssen die angeforderten CPUs der Teilreservierungen exakt festgelegt werden. Die Cluster und die Startzeiten, bei denen die Teilreservierungen durchgeführt werden, ermittelt ein Co-Reservierungs-Algorithmus erst bei der Anfragebearbeitung. Die Benutzer können die Startzeit in der Anfrage eingrenzen, in dem sie eine früheste Startzeit und eine Deadline angeben. In den bisher entwickelten Reservierungsverfahren können Benutzer nur starre Co-Reservierungsanfragen stellen, in denen sie für alle Attribute der Teilreservierungen exakte Werte angeben müssen. Eine flexible Co-Reservierungsanfrage erlaubt einem Grid-Reservierungsdienst optimierte Reservierungsvarianten durchzuführen, da er aus vielen Reservierungsvarianten auswählen kann. Die Anwendungen von lokalen Benutzern können durch die Grid-Reservierungen verzögert werden. Bei der Auswahl der Reservierungsvarianten kann ein Grid-Reservierungsdienst dies berücksichtigen und den negativen Einfluß der Grid-Reservierungen verringern. Bei zwei Co-Reservierungs-Algorithmen werden auch zusätzliche Sortiervorgaben des Klienten bei der Auswahl der Reservierungsvarianten beachtet. Die Algorithmen wurden durch Simulationen mit dem Workload eines realen großen Supercomputers ausgewertet. Die Untersuchungen zeigen, daß sowohl die lokalen als auch die Grid-Benutzer von den flexiblen Anfragen profitieren können. Den lokalen Benutzern wurden die Ressourcen weniger oft durch zu früh gestellte Reservierungsanfragen entzogen. Die Grid- Benutzer können ihrerseits komfortabel eine optimierte Reservierungsvariante durchführen lassen.

Beschreibung: This diploma thesis deals with the restoration problem in telecommunication networks. The goal is to find a cost minimal capacity capacity assignment on the edges and nodes of a network such that given demands can be satisfied even in case of the failure of an edge or node in the network. Moreover, restrictions on the routing paths (like length restrictions) and hardware constraints have to be satisfied. A Mixed Integer Programming model is presented which takes into account restoration requirements as well as hardware constraints and which abstracts from a particular restoration protocol and failure situation. This abstraction provides new insight into the structure of the network restoration problem and shows that from a mathematical point of view, the commonly used restoration techniques Link Restoration, Path Restoration and Reservation are not as different as they seem to be from a practical point of view. In addition, our model allows (but is not limited to) optimizing working capacity, intended for normal use, and spare capacity, intended for rerouting purposes in case of a failure, in one step. Furthermore, our formulation of capacity cost allows taking into account the effects of discrete, non-linear cost structures which are common in practice. Up to our knowledge, no publication in the existing literature covers all these aspects, let alone in one model, although they are of major practical interest. The model has been implemented in a Branch and Cut framework. The theoretical background of the algorithmic procedure is presented in detail, including computational complexity investigations on the pricing problem. The abstraction from a particular restoration protocol turns out to be useful both from a theoretical and computational point of view. In fact, our investigations suggest a distinction into Local Restoration and Global Restoration rather than into Link Restoration,Path Restoration, Reservation and mixtures of these concepts. In addition to the theoretical aspects of the algorithmic procedure, some implementational details are briefly discussed. Our implementation has been tested on 14 real world instances, which is described in detail. One part of the computational results consists of a comparison of optimal network cost values using diffeent restoration mechanisms, applied to securing either all single node failures, all single edge failures or both. In addition, the effects of a discrete cost structure are investigated, which has rarely been considered yet in literature. Furthermore, the cost ifference between joint and successive working and spare capacity optimization is investigated. In the second part of the computational results, several heuristics for the network restoration problem are compared with respect to both solution quality and time. This diploma thesis deals with the restoration problem in telecommunication networks. The goal is to find a cost minimal capacity capacity assignment on the edges and nodes of a network such that given demands can be satisfied even in case of the failure of an edge or node in the network. Moreover, restrictions on the routing paths (like length restrictions) and hardware constraints have to be satisfied. A Mixed Integer Programming model is presented which takes into account restoration requirements as well as hardware constraints and which abstracts from a particular restoration protocol and failure situation. This abstraction provides new insight into the structure of the network restoration problem and shows that from a mathematical point of view, the commonly used restoration techniques Link Restoration, Path Restoration and Reservation are not as different as they seem to be from a practical point of view. In addition, our model allows (but is not limited to) optimizing working capacity, intended for normal use, and spare capacity, intended for rerouting purposes in case of a failure, in one step. Furthermore, our formulation of capacity cost allows taking into account the effects of discrete, non-linear cost structures which are common in practice. Up to our knowledge, no publication in the existing literature covers all these aspects, let alone in one model, although they are of major practical interest. The model has been implemented in a Branch and Cut framework. The theoretical background of the algorithmic procedure is presented in detail, including computational complexity investigations on the pricing problem. The abstraction from a particular restoration protocol turns out to be useful both from a theoretical and computational point of view. In fact, our investigations suggest a distinction into Local Restoration and Global Restoration rather than into Link Restoration, Path Restoration, Reservation and mixtures of these concepts. In addition to the theoretical aspects of the algorithmic procedure, some implementational details are briefly discussed. Our implementation has been tested on 14 real world instances, which is described in detail. One part of the computational results consists of a comparison of optimal network cost values using different restoration mechanisms, applied to securing either all single node failures, all single edge failures or both. In addition, the effects of a discrete cost structure are investigated, which has rarely been considered yet in literature. Furthermore, the cost difference between joint and successive working and spare capacity optimization is investigated. In the second part of the computational results, several heuristics for the network restoration problem are compared with respect to both solution quality and time.