Forecasting and anticipating SLO breaches in programmable networks

Jaafar Bendriss (1, 2, 3, 4) , Imen Grida Ben Yahia (3) , Djamal Zeghlache (5, 1, 2, 4)

1 TSP - RS2M - Département Réseaux et Services Multimédia Mobiles
2 SAMOVAR - Services répartis, Architectures, MOdélisation, Validation, Administration des Réseaux
3 Orange Labs [Chatillon]
4 CNRS - Centre National de la Recherche Scientifique
5 R3S-SAMOVAR - Réseaux, Systèmes, Services, Sécurité

Software Networks built by combining Software Defined Networks (SDN), Network Function Virtualization (NFV) and Cloud principles call for agile and dynamic automation of management operations to ensure continuous provisioning and deployment of networked resources and services. In this context, efficient Service Level Agreements (SLA) management and anticipation of Service Level Objectives (SLO) breaches become essential to fulfill established service contracts with clients. In this paper, we design and specify a framework for cognitive SLA enforcement (using Artificial Neural Network learning) for networking services involving VNFs (Virtualized Network Functions) and SDN controllers. A proof of concept, a testbed description and an extensive evaluation assess the performance of the proposed framework

Mots clés: SDN; NFV; SLO; SLA; ANN; Cognitive management; Network management; Machine learning

Domaines: Réseaux et télécommunications [cs.NI]

Jaafar Bendriss, Imen Grida Ben Yahia, Djamal Zeghlache. Forecasting and anticipating SLO breaches in programmable networks. ICIN 2017 : 20th Conference on Innovations in Clouds, Internet and Networks, Mar 2017, Paris, France. pp.127 - 134, ⟨10.1109/ICIN.2017.7899402⟩. ⟨hal-01531499⟩