Giordano, Ugo (2017) In-Production Continuous Testing for Future Telco Cloud. [Tesi di dottorato]


Download (6MB) | Preview
[error in script] [error in script]
Item Type: Tesi di dottorato
Lingua: English
Title: In-Production Continuous Testing for Future Telco Cloud
Date: 5 October 2017
Institution: Università degli Studi di Napoli Federico II
Department: Ingegneria Elettrica e delle Tecnologie dell'Informazione
Dottorato: Information technology and electrical engineering
Ciclo di dottorato: 29
Coordinatore del Corso di dottorato:
Date: 5 October 2017
Uncontrolled Keywords: Software Defined Network, Failure Injection, SDN, SDN performance
Settori scientifico-disciplinari del MIUR: Area 09 - Ingegneria industriale e dell'informazione > ING-IND/05 - Impianti e sistemi aerospaziali
Date Deposited: 22 Nov 2017 11:44
Last Modified: 08 Mar 2018 13:30
DOI: 10.6093/UNINA/FEDOA/11884


Software Defined Networking (SDN) is an emerging paradigm to design, build and operate networks. The driving motivation of SDN was the need for a major change in network technologies to support a configuration, management, operation, reconfiguration and evolution than in current computer networks. In the SDN world, performance it is not only related to the behaviour of the data plane. As the separation of control plane and data plane makes the latter significantly more agile, it lays off all the complex processing workload to the control plane. This is further exacerbated in distributed network controller, where the control plane is additionally loaded with the state synchronization overhead. Furthermore, the introduction of SDNs technologies has raised advanced challenges in achieving failure resilience, meant as the persistence of service delivery that can justifiably be trusted, when facing changes, and fault tolerance, meant as the ability to avoid service failures in the presence of faults. Therefore, along with the “softwarization” of network services, it is an important goal in the engineering of such services, e.g. SDNs and NFVs, to be able to test and assess the proper functioning not only in emulated conditions before release and deployment, but also “in-production”, when the system is under real operating conditions.   The goal of this thesis is to devise an approach to evaluate not only the performance, but also the effectiveness of the failure detection, and mitigation mechanisms provided by SDN controllers, as well as the capability of the SDNs to ultimately satisfy nonfunctional requirements, especially resiliency, availability, and reliability. The approach consists of exploiting benchmarking techniques, such as the failure injection, to get continuously feedback on the performance as well as capabilities of the SDN services to survive failures, which is of paramount importance to improve the effective- ness of the system internal mechanisms in reacting to anomalous situations potentially occurring in operation, while its services are regularly updated or improved. Within this vision, this dissertation first presents SCP-CLUB (SDN Control Plane CLoUd-based Benchmarking), a benchmarking frame- work designed to automate the characterization of SDN control plane performance, resilience and fault tolerance in telco cloud deployments. The idea is to provide the same level of automation available in deploying NFV function, for the testing of different configuration, using idle cycles of the telco cloud infrastructure. Then, the dissertation proposes an extension of the framework with mechanisms to evaluate the runtime behaviour of a Telco Cloud SDN under (possibly unforeseen) failure conditions, by exploiting the software failure injection.


Downloads per month over past year

Actions (login required)

View Item View Item