Tesi etd-02032025-153718 |
Link copiato negli appunti
Tipo di tesi
Tesi di laurea magistrale
Autore
CASU, FEDERICO
URN
etd-02032025-153718
Titolo
Energy Efficient Resource Management for Cloud Native Network Functions
Dipartimento
INGEGNERIA DELL'INFORMAZIONE
Corso di studi
COMPUTER ENGINEERING
Relatori
relatore Virdis, Antonio
relatore Vallati, Carlo
relatore Vallati, Carlo
Parole chiave
- BIP
- CNF placement
- edge computing
- energy efficiency
- SFC
- SLA
Data inizio appello
21/02/2025
Consultabilità
Non consultabile
Data di rilascio
21/02/2028
Riassunto
Network Function Virtualization (NFV) has emerged as a well established paradigm in modern networking, enabling the decoupling of network functions from proprietary hardware. This architectural shift facilitates the deployment of Virtual Network Functions (VNFs) and Cloud-Native network Functions (CNFs) on commodity hardware, enhancing scalability, flexibility, and cost efficiency. Managing VNF and CNF presents, however, several challenges, particularly when it comes to resource allocation. Moreover, in modern deployments VNFs are instantiated in small or micro datacenters placed at the edge of the network. This introduces an heterogeneity in computing, communication and consumption characteristics of the involved nodes, thus further increasing the complexity of deploying VNFs and CNFs. This thesis contributes the field of NFV by introducing a structured methodology for CNF placement that optimizes energy efficiency while ensuring compliance with Service Level Agreements (SLAs). Specifically, it focuses on the strategic allocation of CNFs within NFV-enabled edge infrastructures by consolidating CNF instances on energy-efficient servers. High-performance, power-intensive servers are utilized only when necessary, such as in scenarios requiring specialized hardware accelerators to meet specific computational demands. The proposed CNF placement algorithm is formulated as a Binary Integer Programming (BIP) optimization problem, enabling precise decision-making in resource allocation while minimizing energy consumption. To evaluate the effectiveness of the proposed approach, the BIP model has been evaluated against heuristic-based solutions across diverse workload scenarios. This assessment aims to determine whether the proposed solution achieves more energy-efficient allocation schemes while maintaining compliance with critical SLAs, such as ensuring maximum end-to-end delay constraints.
File
Nome file | Dimensione |
---|---|
La tesi non è consultabile. |