Dissertations / Theses on the topic 'Placement des VNF'

The evolution of information and communication technologies (e.g. cloud computing, the Internet of Things (IoT) and 5G, among others) has enabled a large market of applications and network services for a massive number of users connected to the Internet. Achieving high programmability while decreasing complexity and costs has become an essential aim of networking research due to the ever-increasing pressure generated by these applications and services. However, meeting these goals is an almost impossible task using traditional IP networks. Software-Defined Networking (SDN) is an emerging network architecture that could address the needs of service providers and network operators. This new technology consists in decoupling the control plane from the data plane, enabling the centralization of control functions on a concentrated or distributed platform. It also creates an abstraction between the network infrastructure and network applications, which allows for designing more flexible and programmable networks. Recent trends of increased user demands, the explosion of Internet traffic and diverse service requirements have further driven the interest in the potential capabilities of SDN to enable the introduction of new protocols and traffic management models. This doctoral research is focused on improving high-level policies and control strategies, which are becoming increasingly important given the limitations of current solutions for large-scale SDN environments. Specifically, the three largest challenges addressed in the development of this thesis are related to the processes of topology discovery, fault recovery and Virtual Network Function (VNF) placement in software-defined and virtualized networks. These challenges led to the design of a set of effective techniques, ranging from network protocols to optimal and heuristic algorithms, intended to solve existing problems and contribute to the deployment and adoption of such programmable networks.For the first challenge, this work presents a novel protocol that, unlike existing approaches, enables a distributed layer 2 discovery without the need for previous IP configurations or controller knowledge of the network. By using this mechanism, the SDN controller can discover the network view without incurring scalability issues, while taking advantage of the shortest control paths toward each switch. Moreover, this novel approach achieves noticeable improvement with respect to state-of-the-art techniques. To address the resilience concern of SDN, we propose a self-healing mechanism that recovers the control plane connectivity in SDN-managed environments without overburdening the controller performance. The main idea underlying this proposal is to enable real-time recovery of control paths in the face of failures without the intervention of a controller. Obtained results show that the proposed approach recovers the control topology efficiently in terms of time and message load over a wide range of generated networks. The third contribution made in this thesis combines topology knowledge with bin packing techniques in order to efficiently place the required VNF. An online heuristic algorithm with low-complexity was developed as a suitable solution for dynamic infrastructures. Extensive simulations, using network topologies representative of different scales, validate the good performance of the proposed approaches regarding the number of required instances and the delay among deployed functions. Additionally, the proposed heuristic algorithm improves the execution times by a fifth order of magnitude compared to the optimal formulation of this problem.

L’objectif de cette thèse est de présenter de nouveaux algorithmes de répartition des ressources sous la forme de machines virtuelles (VMs) et fonction de réseau virtuel (VNFs) dans les Clouds et réseaux de diffusion de contenu (CDNs). La thèse comprend deux principales parties: la première se concentre sur la rentabilité des Clouds distribués, et développe ensuite les raisons d’optimiser les coûts ainsi que les émissions de carbone. Cette partie comprend quatre contributions. La première contribution est une étude de l’état de l’art sur la répartition des coûts et des émissions de carbone dans les environnements de clouds distribués. La deuxième contribution propose une méthode d’allocation des ressources, appelée NACER, pour les clouds distribués. La troisième contribution présente une méthode de placement VM efficace en termes de coûts et de carbone (appelée CACEV) pour les clouds distribués verts. Pour obtenir une meilleure performance, la quatrième contribution propose une méthode dynamique de placement VM (D-CACEV) pour les clouds distribués. La deuxième partie propose des algorithmes de placement de VNFs dans les Clouds et réseaux de CDNs pour optimiser les coûts. Cette partie comprend cinq contributions. Une étude de l’état de l’art sur les solutions proposées est le but de la première contribition. La deuxième contribution propose une méthode d’allocation des ressources, appelée CCVP, pour le provisionnement de service réseau dans les clouds et réseaux de ISP. La troisième contribution implémente le résultat de l’algorithme CCVP dans une plateforme réelle. La quatrième contribution considère l’effet de la permutation de VNFs dans les chaîne de services et la cinquième contribution explique le placement de VNFs pour les services à valeur ajoutée dans les CDNs
High energy costs and carbon emissions are two significant problems in distributed computing domain, such as distributed clouds and Content Delivery Networks (CDNs). Resource allocation methods (e.g., in form of Virtual Machine (VM) or Virtual Network Function (VNF) placement algorithms) have a direct effect on cost, carbon emission and Quality of Service (QoS). This thesis includes three related parts. First, it targets the problem of resource allocation (i.e., in the form of network aware VM placement algorithms) for distributed clouds and proposes cost and carbon emission efficient resource allocation algorithms for green distributed clouds. Due to the similarity of the network-aware VM placement problem in distributed clouds with a VNF placement problem, the second part of the thesis, getting experience from the first part, proposes a new cost efficient resource allocation algorithm (i.e., VNF placement) for network service provision in data centers and Internet Service Provider (ISP) network. Finally, the last part of the thesis presents new cost efficient resource allocation algorithms (i.e., VNF placement) for value-added service provisioning in NFV-based CDNs

Le but de cette thèse est d’étudier et d’évaluer le rôle de la virtualisation des réseau de diffusion de contenu. Nous proposons une implémentation d’une architecture CDN permettant à un opérateur de réseau de virtualiser son infrastructure CDN et de la louer à des fournisseurs de contenu. Afin d’avoir une allocation optimale des ressources, nous proposons une méthode qui combine les informations fournies lors de la demande par le fournisseur de contenu avec les données du réseau et de l’infrastructure de calcul. Nous avons modélisé ce problème d’allocation de ressources en problème d’optimisation, résolu par un algorithme. Les résultats obtenues donnent suite à la proposition d’algorithmes et d’heuristiques de placement pour l’allocation conjointe de vCPU-à-VM et le placement des VMs dans les Pms
The goal of this thesis is to study and evaluate the role a Virtual CDNs in improving the end-users QoE while saving on service providers’ costs and service availability. First, we present the design and implementation of an architecture for on-demand deployment of a vCDN infrastructure over a telco cloud. Second, we propose different algorithms for solving the Virtual Network Function (VNF) placement problem. We propose a polynomialtime heuristic algorithms to solve a relaxed version of the problem’s assumptions, we show experimentally that the derived solutions are close to the optimal. Finally, we study and evaluate solutions for the placement of VNF at the edge, by moving from the traditional central cloud to the edge one. We have also shown how our method can reduce delays and still provide a highly-available service

5G and beyond mobile network technology promises to deliver unprecedented ultra-low latency and high data rates, paving the way for many novel applications and services. Network Function Virtualization (NFV) and Multi-access Edge Computing (MEC) are two technologies expected to play a vital role in achieving ambitious Quality of Service requirements of such applications. While NFV provides flexibility by enabling network functions to be dynamically deployed and inter-connected to realize Service Function Chains (SFC), MEC brings the computing capability to the mobile network's edges, thus reducing latency and alleviating the transport network load. However, adequate mechanisms are needed to meet the dynamically changing network service demands (i.e., in single and multiple domains) and optimally utilize the network resources while ensuring that the end-to-end latency requirement of services is always satisfied. In this dissertation work, we break the problem into three separate stages and present the solutions for each one of them.Firstly, we apply Artificial Intelligence (AI) techniques to drive NFV resource orchestration in MEC-enabled 5G architectures for single and multi-domain scenarios. We propose three deep learning approaches to perform horizontal and vertical Virtual Network Function (VNF) auto-scaling: (i) Multilayer Perceptron (MLP) classification and regression (single-domain), (ii) Centralized Artificial Neural Network (ANN), centralized Long-Short Term Memory (LSTM) and centralized Convolutional Neural Network-LSTM (CNN-LSTM) (single-domain), and (iii) Federated ANN, federated LSTM and federated CNN-LSTM (multi-domain). We evaluate the performance of each of these deep learning models trained over a commercial network operator dataset and investigate the pros and cons of different approaches for VNF auto-scaling. For the first approach, our results show that both MLP classifier and MLP regressor models have strong predicting capability for auto-scaling. However, MLP regressor outperforms MLP classifier in terms of accuracy. For the second approach (one-step prediction), CNN-LSTM performs the best for the QoS-prioritized objective and LSTM performs the best for the cost-prioritized objective. For the second approach (multi-step prediction), the encoder-decoder CNN-LSTM model outperforms the encoder-decoder LSTM model for both QoS and Cost prioritized objectives. For the third approach, both federated LSTM and federated CNN-LSTM models perform equally better than the federated ANN model. It was also noted that in general federated learning approaches performs poorly compared to centralized learning approaches. Secondly, we employ Integer Linear Programming (ILP) techniques to formulate and solve a joint user association and SFC placement problem, where each SFC represents a service requested by a user with end-to-end latency and data rate requirements. We also develop a comprehensive end-to-end latency model considering radio delay, backhaul network delay and SFC processing delay for 5G mobile networks. We evaluated the proposed model using simulations based on real-operator network topology and real-world latency values. Our results show that the average end-to-end latency reduces significantly when SFCs are placed at the ME hosts according to their latency and data rate demands. Furthermore, we propose an heuristic algorithm to address the issue of scalability in ILP, that can solve the above association/mapping problem in seconds rather than hours.Finally, we introduce lightMEC - a lightweight MEC platform for deploying mobile edge computing functionalities which allows hosting of low-latency and bandwidth-intensive applications at the network edge. Measurements conducted over a real-life test demonstrated that lightMEC could actually support practical MEC applications without requiring any change to existing mobile network nodes' functionality in the access and core network segments. The significant benefits of adopting the proposed architecture are analyzed based on a proof-of-concept demonstration of the content caching use case. Furthermore, we introduce the AI-driven Kubernetes orchestration prototype that we implemented by leveraging the lightMEC platform and assess the performance of the proposed deep learning models (from stage 1) in an experimental setup. The prototype evaluations confirm the simulation results achieved in stage 1 of the thesis.

L’objectif de cette thèse est de présenter de nouveaux algorithmes de répartition des ressources sous la forme de machines virtuelles (VMs) et fonction de réseau virtuel (VNFs) dans les Clouds et réseaux de diffusion de contenu (CDNs). La thèse comprend deux principales parties: la première se concentre sur la rentabilité des Clouds distribués, et développe ensuite les raisons d’optimiser les coûts ainsi que les émissions de carbone. Cette partie comprend quatre contributions. La première contribution est une étude de l’état de l’art sur la répartition des coûts et des émissions de carbone dans les environnements de clouds distribués. La deuxième contribution propose une méthode d’allocation des ressources, appelée NACER, pour les clouds distribués. La troisième contribution présente une méthode de placement VM efficace en termes de coûts et de carbone (appelée CACEV) pour les clouds distribués verts. Pour obtenir une meilleure performance, la quatrième contribution propose une méthode dynamique de placement VM (D-CACEV) pour les clouds distribués. La deuxième partie propose des algorithmes de placement de VNFs dans les Clouds et réseaux de CDNs pour optimiser les coûts. Cette partie comprend cinq contributions. Une étude de l’état de l’art sur les solutions proposées est le but de la première contribition. La deuxième contribution propose une méthode d’allocation des ressources, appelée CCVP, pour le provisionnement de service réseau dans les clouds et réseaux de ISP. La troisième contribution implémente le résultat de l’algorithme CCVP dans une plateforme réelle. La quatrième contribution considère l’effet de la permutation de VNFs dans les chaîne de services et la cinquième contribution explique le placement de VNFs pour les services à valeur ajoutée dans les CDNs
High energy costs and carbon emissions are two significant problems in distributed computing domain, such as distributed clouds and Content Delivery Networks (CDNs). Resource allocation methods (e.g., in form of Virtual Machine (VM) or Virtual Network Function (VNF) placement algorithms) have a direct effect on cost, carbon emission and Quality of Service (QoS). This thesis includes three related parts. First, it targets the problem of resource allocation (i.e., in the form of network aware VM placement algorithms) for distributed clouds and proposes cost and carbon emission efficient resource allocation algorithms for green distributed clouds. Due to the similarity of the network-aware VM placement problem in distributed clouds with a VNF placement problem, the second part of the thesis, getting experience from the first part, proposes a new cost efficient resource allocation algorithm (i.e., VNF placement) for network service provision in data centers and Internet Service Provider (ISP) network. Finally, the last part of the thesis presents new cost efficient resource allocation algorithms (i.e., VNF placement) for value-added service provisioning in NFV-based CDNs

Cette thèse présente un cadre flexible basé sur l'apprentissage par renforcement des files d'attente pour l'orchestration dynamique des tranches dans les réseaux Beyond 5G, prenant en charge de multiples tranches concurrentes qui couvrent différents domaines technologiques et sont régies par divers accords de niveau de service de bout en bout. Différentes méthodes d'apprentissage par renforcement profond (mono ou multi-agents) sont étudiées pour résoudre les problèmes de complexité d'état et d'action liés à ces problèmes combinatoires, qui rendent l'utilisation d'algorithmes d'apprentissage par renforcement classique impraticable. La performance des schémas proposés est validée par des simulations dans des scénarios de trafic markovien synthétique et de trafic réel
This Thesis introduces a flexible Reinforcement Learning queuing-based framework for dynamic slice orchestration in Beyond 5G networks, supporting multiple concurrent slices that span different technological domains and are governed by diverse end-to-end Service Level Agreements. Different (Deep) Reinforcement Learning methods (single or multi-agent) are investigated to address the state and action complexity hurdles arising in such combinatorial problems, which render the use of "vanilla" Reinforcement Learning algorithms impractical. The performance of the proposed schemes is validated through simulations under both synthetic Markovian traffic and real traffic scenarios

NFV is an emerging network architecture to increase flexibility and agility within operator's networks by placing virtualized services on demand in Cloud data centers (CDCs). One of the main challenges for the NFV environment is how to efficiently allocate Virtual Network Functions (VNFs) to Virtual Machines (VMs) and how to minimize network latency in the rapidly changing network environments. Although a significant amount of work/research has been already conducted for the generic VNF placement problem and VM migration for efficient resource management in CDCs, network latency among various network components and VNF migration problem have not been comprehensively considered yet to the best of our knowledge. Firstly, to address VNF placement problem, we design a more comprehensive model based on real measurements to capture network latency among VNFs with more granularity to optimize placement of VNFs in CDCs. We consider resource demand of VNFs, resource capacity of VMs and network latency among various network components. Our objectives are to minimize both network latency and lead time (the time to find a VM to host a VNF). Experimental results are promising and indicate that our approach, namely VNF Low-Latency Placement (VNF-LLP), can reduce network latency by up to 64.24% compared with two generic algorithms. Furthermore, it has a lower lead time as compared with the VNF Best-Fit Placement algorithm. Secondly, to address VNF migration problem, we i) formulate the VNF migration problem and ii) develop a novel VNF migration algorithm called VNF Real-time Migration (VNF-RM) for lower network latency in dynamically changing resource availability. As a result of experiments, the effectiveness of our algorithm is demonstrated by reducing network latency by up to 59.45% after latency-aware VNF migrations.

Comme le Wi-Fi est devenu de plus en plus important dans les réseaux actuels, ainsi que dans les réseaux du futur, de nouvelles exigences " opérateur " se sont apparues afin de supporter les attentes des utilisateurs et de fournir des réseaux Wi-Fi de haute performance. Dans ce contexte, nous étudions plusieurs problèmes liés à la conception et l'optimisation des réseaux Wi-Fi opérateur de nouvelle génération. Dans la première étape, notre objectif est d'améliorer l'expérience utilisateur Wi-Fi et de lui offrir un accès personnalisé et transparent aux réseaux et services Wi-Fi. Pour cela, nous proposons une extension des trames de gestion IEEE 802.11 pour activer la découverte des services locaux avant l'association Wi-Fi, tout en évitant la surcharge du canal. Nous définissons également un ensemble d'étiquettes de service pour identifier d'une manière standardisée les services les plus connus. Dans la deuxième étape, nous adressons les problèmes liés à l'architecture et la gestion du réseau dans un environnement Wi-Fi opérateur de nouvelle génération. Plus précisément, nous proposons, tout d'abord, une nouvelle architecture Wi-Fi qui exploite les concepts de NFV et du Edge Cloud Computing. Nous visons à travers cette architecture à apporter plus d'agilité et d'adaptabilité et d'améliorer la QoS perçue par l'utilisateur en plaçant des fonctions réseau et certains services à proximité de lui. Pour faire face à certains problèmes de gestion dans cette architecture, nous proposons ensuite des stratégies de placement et de provisionnement des fonctions de réseau virtuelles en s'appuyant sur des exigences de QoS
As Wi-Fi is gaining a lot of momentum in today’s networks as well as in future networks, new carrier-grade requirements are emerging to support future user expectations and provide high-performance Wi-Fi networks. In this context, we investigate several problems surrounding the design and optimization of carrier-grade next-generation Wi-Fi networks. In the first stage, our objective is to improve the Wi-Fi user experience and offer to him a personalized and seamless access to Wi-Fi networks and services. For this, we propose an extension to the IEEE 802.11 management frames to enable venue service discovery prior to Wi-Fi association while avoiding channel overhead. We define also a set of extensible service labels to uniquely and globally identify the most known venue-based services. In the second stage, we deal with network architecture and management issues in next-generation carrier Wi-Fi environment. More specifically, we first propose a novel carrier-managed Wi-Fi architecture that leverages NFV and Edge Cloud Computing concepts. We aim through this architecture to bring more agility and adaptability and improve user perceived QoS by placing network functions and certain services close to end-users. To address some major management issues in this proposed architecture, we then propose placement and provisioning strategies of Virtual Network Functions based on QoS requirements. These strategies can also be applied to any edge-central wireless carrier architecture, since they do not make any assumption about the underlying wireless technology

Comme le Wi-Fi est devenu de plus en plus important dans les réseaux actuels, ainsi que dans les réseaux du futur, de nouvelles exigences " opérateur " se sont apparues afin de supporter les attentes des utilisateurs et de fournir des réseaux Wi-Fi de haute performance. Dans ce contexte, nous étudions plusieurs problèmes liés à la conception et l'optimisation des réseaux Wi-Fi opérateur de nouvelle génération. Dans la première étape, notre objectif est d'améliorer l'expérience utilisateur Wi-Fi et de lui offrir un accès personnalisé et transparent aux réseaux et services Wi-Fi. Pour cela, nous proposons une extension des trames de gestion IEEE 802.11 pour activer la découverte des services locaux avant l'association Wi-Fi, tout en évitant la surcharge du canal. Nous définissons également un ensemble d'étiquettes de service pour identifier d'une manière standardisée les services les plus connus. Dans la deuxième étape, nous adressons les problèmes liés à l'architecture et la gestion du réseau dans un environnement Wi-Fi opérateur de nouvelle génération. Plus précisément, nous proposons, tout d'abord, une nouvelle architecture Wi-Fi qui exploite les concepts de NFV et du Edge Cloud Computing. Nous visons à travers cette architecture à apporter plus d'agilité et d'adaptabilité et d'améliorer la QoS perçue par l'utilisateur en plaçant des fonctions réseau et certains services à proximité de lui. Pour faire face à certains problèmes de gestion dans cette architecture, nous proposons ensuite des stratégies de placement et de provisionnement des fonctions de réseau virtuelles en s'appuyant sur des exigences de QoS
As Wi-Fi is gaining a lot of momentum in today’s networks as well as in future networks, new carrier-grade requirements are emerging to support future user expectations and provide high-performance Wi-Fi networks. In this context, we investigate several problems surrounding the design and optimization of carrier-grade next-generation Wi-Fi networks. In the first stage, our objective is to improve the Wi-Fi user experience and offer to him a personalized and seamless access to Wi-Fi networks and services. For this, we propose an extension to the IEEE 802.11 management frames to enable venue service discovery prior to Wi-Fi association while avoiding channel overhead. We define also a set of extensible service labels to uniquely and globally identify the most known venue-based services. In the second stage, we deal with network architecture and management issues in next-generation carrier Wi-Fi environment. More specifically, we first propose a novel carrier-managed Wi-Fi architecture that leverages NFV and Edge Cloud Computing concepts. We aim through this architecture to bring more agility and adaptability and improve user perceived QoS by placing network functions and certain services close to end-users. To address some major management issues in this proposed architecture, we then propose placement and provisioning strategies of Virtual Network Functions based on QoS requirements. These strategies can also be applied to any edge-central wireless carrier architecture, since they do not make any assumption about the underlying wireless technology

碩士
國立交通大學
資訊科學與工程研究所
107
Over the last few years, network function virtualization (NFV) has been developing rapidly in both academia and industry since it promises flexible and scalable network management and orchestration. The European Telecommunications Standards Institute (ETSI) has designed NFV management and orchestration (NFV-MANO) architectural framework for the purpose of managing and orchestrating the network services and their corresponding virtualized network functions (VNFs). In the NFV-MANO, the NFV orchestrator (NFVO) manages the network services and coordinates with the VNF managers (VNFMs) which manage the lifecycle of VNFs. It is a challenging problem to determine the placement of VNFMs that keeps the operational cost down without violating the performance requirements. Existing centralized solution that suffers from scalability and privacy issues, in this paper, we leverage potential game theory to achieve distributed solutions to the VNFM placement problem (MPP). More importantly, our distributed solutions hold the property of self-adaptability. It can adapt the changes in the system like some VNFMs are faulty or the number of VNF instances varies. On the other hand, we also show that the game possesses a potential function and admits a Nash equilibrium (NE). Numerical results show the cost efficiency and adaptability of our approaches.

碩士
國立中興大學
資訊科學與工程學系所
107
With new technological advances, application services have become more and more diversified. The applications of Augmented Reality (AR) have received great attention from academics and operators. However, the complex computing required for AR application services has caused service operators to face problems of poor user experience due to insufficient computing resources. Network function virtualization (NFV) enables operators to directly establish network functions on virtual machines, and can flexibly adjust the number of each virtual network function (VNF) according to users’ requirements. Thus, the system can reduce unnecessary waste of resources. In addition, we use multi-access edge computing (MEC) to provide computing resources. The edge server is close to the user and reduces transmission cost of the system. This paper proposes an MEC-based VNF placement and scheduling scheme for AR application. Considering the function topology of AR application, the compute-intensive application can provide good service quality for a large number of users within limited resources. The simulation presents that the proposed method can maintain good service quality as the number of user requests increases. Moreover, we use queueing theory to design a queueing model for the MEC-based VNF scheduling algorithm and use mathematical analysis to validate simulation results.

碩士
國立中正大學
通訊工程研究所
105
Network Function Virtualization (NFV) as a new policy enforcement and service provisioning paradigm is considered as a framework for handling VNF placement and orchestration in the network. Each incoming request as a sequence of Network Functions (NFs), called Service Chain (SC), is processed and the environment is optimally prepared for satisfying the request. The SC provisioning has some operational costs. Creating VNF for each VNF request increases the number of VNFs in an unpleasant manner that leads to the higher resource wastage, environmental complexity and management costs. Because of the improper placement of the VNFs for each SC, the bandwidth usage cost affects the performance of the network. For decreasing the VNF numbers and bandwidth usage costs in the network, we propose Central Orchestration System (COS) which integrates the benefits of OpenStack in the presence of SDN technology. Our proposed method, according to the user request and the current situation of the network, calculates the optimum VNF locations and the connecting path of the VNFs in order to minimize the bandwidth and number of VNF costs. The first experiment confirms the computation time of our optimization model in a 48-ary Fat tree network environment. In the second experiment the COS had the better performance in comparison with the other methods.

The Smart Home concept refers to a domestic network environment able to provide connectivity, adaptivity and automation through different kinds of smart device to improve inhabitants’ awareness, safety and comfort. This thesis is concerned with the evolution of the Smart Home Environment towards a new virtual version by mainly dealing with two topics. First, on the application perspective, focus is made on the crucial issue of improving user awareness on power consumptions for achieving energy savings. Second, on the technological infrastructure perspective, it investigates possible impact and related issues of the adoption of network virtualization and programmability technologies in the Smart Home networking by providing a contribution as regards VNF Placement problem. The rational use of energy has recently become one of the most pressing research topic because of the constantly growing consumptions in contrast with the scarcity of resources. In a Smart Home scenario the recent progress of technology, along with lower costs, has made it possible to perform energy monitoring and management actions through the distribution of smart meters and environmental sensors capable of providing information to a Home Energy Management System (HEMS). Recent studies have shown that informing users about the actual appliances consumption as well as device-usage habits, can help to obtain energy consumption reduction in private households. In order to achieve this goal a supervised classification algorithm for detecting and identifying consuming appliances has been implemented. Then a Non-Intrusive Load Monitoring (NILM) approach has been investigated to reduce the cost of attaching a single meter (i.e. smart plug) to each device; the proposed algorithm aims at recognizing the power consumption of a specific device from the whole-house consumption profile and from the input of context information (i.e. the user presence in the house and the hourly utilization of appliances). As regards the technological perspective, the Network Function Virtualization (NFV) technology, together with the complementary Software Defined Networking (SDN) paradigm, let envisage a revolution in the traditional concept of network service delivery and availability. This evolution promises to enable on-demand and flexible services provision as it allows to separate the network functions from the hardware they run on by leveraging the virtualization abstraction. This work addresses the issue of optimal VNF placement in a multi-stakeholder network infrastructure by considering the framework of a NFV Management and Orchestration architecture that leverages the Software Defined Networking paradigm. Given a set of service requests and considering a set of constraints (e.g., maximum end-to-end delay, monetary cost, allowed server utilization level), a mathematical model has been formulated to maximize the profit that can be obtained by both tenants (i.e. Infrastructure providers, Cloud providers) and renters (i.e. service providers/users). In order to favour generalization and to ease the treatment of aspects that in literature have not accounted (e.g. multiple users), the choice of the actual forwarding path of incoming traffic flows is deferred to a later step (optimal routing), to be performed by the SDN Controller. Moreover, the work provides a detailed formalization of service requests and Data Centers and considers two types of users with different privileges (i.e. Premium and Best Effort). The energy efficiency and sustainability goals have been also taken into account.

碩士
國立中正大學
通訊工程研究所
107
Network Function Virtualization (NFV)is a technology that implement traditional physical network devices into virtual machine placed in the server. The virtual machine(VM) with network function is called Virtual Network Function(VNF), users can specify a service chain formed by multiple VNFs in a specific order, this is called Service Chain(SC). If every time a user apply for a Service Chain in the data center and the manager create new VNFs for it this will cause too many VNFs in the environment and increase the management cost and the waste of resource in the server. Randomly placing the VNFs in the environment will cause the routing path length of Service Chain increase and the waste of network bandwidth. We implement a System Orchestrtor (SO). When getting a SC request from the tenant, manager only need to input the SC information, such as VNF type, the order of VNFs, and the bandwidth require of the SC, SO can automatally deploy VNFs in the OpenStack environment based on the CPU usage of VNF, sharing status of the VNF, CPU and RAM usage of server, and network staus. Plus SO will add SC routing entry on P4 switch.

Telecom Service Providers (TSPs) were traditionally dependent on physical devices to provide end-to-end communication. The services provided were high quality and stable but low in agility and hardware-dependent. As the demand for quick deployment of diverse services increased, TSP-s needed much higher flexibility and agility. This is how Network Functions Virtualization (NFV) came into being. NFV is the concept of replacing dedicated hardware with commercial-off-the-shelf (COTS) servers. It decouples the physical hardware and the function running on it. A network function can be dispatched as an instance of the software, called a Virtual Network Function (VNF). Thus, a service can be decomposed into several VNFs that can be run on industry-standard physical servers. The optimal placement of these VNFs is a potential question for TSPs to reduce the overall cost. We first study a network operations problem where we optimally deploy VNFs in Service Chains (SCs) such that the maximum consumed bandwidth across network links is minimized. The network parameters (link bandwidths, compute capacities of nodes, link propagation delays, etc.) and the number of SCs are known a priori. The problem formulated is a large Mixed-Integer Linear Program (MILP). We use the Column Generation (CG) technique to solve the problem optimally. Through various examples, we show the power of CG. We compare our results with recent heuristics and demonstrate that our approach performs better as it gives exact optimal solutions quickly. Second, we extend our previous setup to the online case where the number of SCs is not known a priori. We serve SC requests as they come. A new SC is implemented on the "residual network" while the previously deployed SCs are undisturbed. The problem formulated is a large MILP, and we use CG as the solution technique. The results show the percentage improvement in the solutions over those obtained using heuristics. Next, we study a network design problem in an Edge Computing Environment. A general communication network has a single Data Center (DC) in its "core," which serves as a gateway to the Internet. For delay-constrained services of the kind needed by online gaming, this model does not suffice because the propagation delay between the subscriber and the DC may be too high. This requires some servers to be located close to the network edge. Thus, the question of the optimal placement of these edge servers arises. To lower the network design cost, it is also essential to ensure good traffic routing, so that aggregate traffic on each link remains as low as possible. This enables lower capacity assignment on each link and thereby minimizes design cost. We study a novel joint optimization problem of network design cost minimization. Edge server placement cost and link capacity assignment cost constitute the total cost. The problem formulated is a large MILP, and we again use CG to solve it. We compare our results with many heuristics and show the improvement in design cost. Finally, we extend the above work by relaxing some assumptions and constraints. Unlike previously, we consider servers with different capacities. Also, a server can serve more than one request depending on its core capabilities. We also consider the split-and-merge of an SC through various paths in the network. The formulated problem can also provide the minimum number of servers to be used. Again, the formulation is a large MILP, and CG is used to solve it exactly.