Secure resource management in Next Generation Networks (NGNs)

Abstract

A new era in the computer networks has begun with the advent of 5G which promises to support demanding applications by facilitating high-performance numbers. Ultra-high data rates, high availability, ultra-low latency and wide area coverage are the main characteristics of modern networks which aim to ensure the high QoE level of the end-user. The Network Providers (NPs) are responsible for the deployment and maintenance of modern networks’ infrastructure which consists of novel hardware devices and software components developed in the last few years under the umbrella of 5G. As the infrastructure becomes more sophisticated and expensive, the NPs seek to maximize the efficiency of its utilization by flexibly sharing resources. Therefore, NPs should cooperate and perform resource management processes in this multivendor ecosystem to guarantee the network requirements needed for the support of demanding applications. This introduces the potential of formation of a new marketplace where competitive NPs (with no established trust) trade resources. Hence, several security related questions arise such as: how secure could be this marketplace that consists of various competitive NPs since no one trusts each other? How could we form a network of trust among participants in this vast competitive market without the existence of a trusted third party in order to avoid the drawbacks of centralized approaches? How can we achieve the desired level of security in networks with highly automated management functionalities? In this thesis, we answer to these questions by studying the use of emerging decentralized technologies in modern networks in order to facilitate the secure resource management in Next Generation Networks (NGNs). To this end, a comprehensive study of blockchain technology takes place, where we examine the characteristics of this hyped technology, and we discuss its main advantages and disadvantages. Furthermore, we discuss the major advancements of modern networks focusing on the software developments (i.e., MANOs, VNFs) and we present the standardization effort of the ETSI with the Zero touch networks and Service Management (ZSM) framework. ZSM aims to enable self-management and self-optimization of modern multivendor networks. The goal is to reduce or even eliminate the human intervention in the network management processes of modern networks in order to avoid human errors and delays. However, this framework presents significant security issues mentioned by the standardization team and examined in this study. In this framework, we explored whether blockchain technology can be the answer to the ZSM’s security issues which automatically unveils a novel research area where the blockchain technology and the ZSM framework are combined to provide secure and dynamic resource management in NGNs. To check the feasibility of implementing a blockchain solution to enable NPs share resources, we designed and implemented such an approach. Apart from the qualitative assessment, to check the performance of the blockchain network, we perform experiments in real testbeds by implementing a private/consortium permissioned blockchain network. Our experiments are conducted in two phases in order to check how the network behaves when certain parameters are changing (number of nodes, SC’s structure, consensus mechanism). The metrics we focus on are: the network’s throughput (number of transactions per second), the latency (time needed for transaction’s validation) and success rate (transactions that have been successfully validated). The results of our experiments are encouraging and motivate us to continue our study and design the architecture of a novel blockchain-based ZSM approach. This architecture presents how blockchain can be integrated with ZSM and address the security issues highlighted by the ZSM standardization team. To provide guidance to prospective adopters of the proposed approach, we identify the most suitable blockchain solution for this particular use case and we also proceed to the definition of certain requirements that the Distributed Ledger Technology (DLT) should fulfil so that emerging solutions can be easily assessed. These requirements draw the profile of the ideal DLT that can be adopted to implement the secure resource management scenario and harness the benefits of both DLTs and ZSM framework without jeopardizing the proper operation of NGNs. Therefore, beyond the study of blockchain technology we examine the Directed Acyclic Graphs which is another promising DLT that presents high performance numbers and is considered more scalable than typical blockchains. At the end of this thesis, we performed a qualitative assessment of existing blockchain and DAG solutions in order to compare them with the characteristics that the ideal solution should present. The result of this assessment showed that none of the existing solutions is the ideal one although each one of them contains at least one valuable characteristic.