Titulo Estágio
Self-organising engine for the Cloud-to-Edge continuum
Áreas de especialidade
Comunicações, Serviços e Infraestruturas
Local do Estágio
IPNLIS
Enquadramento
This master thesis is framed within the scope of the OREOS project (https://www.cisuc.uc.pt/en/projects/oreos-orchestration-and-resource-optimization-for-reliable-and-low-latency-services).
In a Smart Environment, a set of devices in a delimited area oversees the gathering of data processed and analysed by mechanisms to gain insights that allow making educated decisions. In the Cloud-to-Edge continuum, there are a plethora of heterogeneous devices that are prone to performance degradation or even failure. In such a dense environment, zero-touch recovery techniques must prevail over manually configured solutions. The monitoring of the assets can help detect performance drops that trigger self-organising solutions, speeding the recovery of the services and preventing any disruption of Service Level Agreements (SLA).
The research of this work is framed in the scenario previously described, where it is critical to monitor all the assets present in the Smart Environment, namely, physical and virtual devices and services and applications, and self-organise them to guarantee the commitment of SLAs.
References:
- Faieq, S., Front, A., Saidi, R. et al. A context-aware recommendation-based system for service composition in smart environments. SOCA 13, 341–355 (2019).
- Jorge Gallego-Madrid, Ramon Sanchez-Iborra, Pedro M. Ruiz, Antonio F. Skarmeta, Machine learning-based zero-touch network and service management: a survey, Digital Communications and Networks, Volume 8, Issue 2, 2022, Pages 105-123, ISSN 2352-8648.
- de Sousa, N.F.S., Islam, M.T., Mustafa, R.U. et al. Machine Learning-Assisted Closed-Control Loops for Beyond 5G Multi-Domain Zero-Touch Networks. J Netw Syst Manage 30, 46 (2022).
Objetivo
The work will consist of designing and developing a Machine Learning (ML)-based End-to-End self-organising engine for Service Chain Functions (SCF) to be deployed in a Cloud-to-Edge scenario, which must be configured in a controlled testbed. The engine will be fed with monitoring data of multiple domains (e.g., cluster performance, resource consumption) to trigger educated self-organising actions (e.g., Virtual Functions (VF) replicas instantiation) using a zero-touch approach via ML techniques.
The activities performed in this work are framed in the Orchestration and Resource optimisation for rEliable and lOw-latency Services (OREOS) project.
Plano de Trabalhos - Semestre 1
Phase 1: Revision of the state-of-the-art on self-organising and ML techniques applied to the Cloud-to-Edge continuum (15/09/2022 – 30/09/2022)
Phase 2: Study and evaluation of SFC and micro-services applied to the Cloud-to-Edge continuum (01/10/2022 – 15/10/2022)
Phase 3: Identify technological and service requirements for self-organising in the Cloud-to-Edge environments, complying with the zero-touch approach (15/10/2022 – 22/10/2022)
Phase 4: Installation and configuration of the Cloud-to-Edge cluster managing tool (e.g., minikube) to deploy SFCs with container support (23/10/2022 – 07/11/2022)
Phase 5: Installation and configuration of the Cloud-to-Edge monitor tool (e.g., Prometheus) (08/11/2022 – 21/12/2022)
Phase 6: Select and implement a baseline algorithm from the SoA for the End-to-End self-organising engine (e.g., greedy algorithm) (22/12/2022 – 31/12/2022)
Phase 7: Prepare midterm defence document (01/10/2022 - 15/01/2023)
Plano de Trabalhos - Semestre 2
Phase 8: Study and review the replication mechanisms available in the cluster managing tool (01/02/2023 - 15/02/2023)
Phase 9: Design and develop the replication mechanism to use in the End-to-End self-organising engine (16/02/2023 – 28/02/2023)
Phase 10: Design and implement at least one algorithm for the educated selection of replicas to be used in the End-to-End self-organising engine (01/03/2023 - 31/05/2023)
Phase 11: Evaluate the performance of the End-to-End self-organising engine in the use case defined (01/06/2023 - 15/06/2023)
Phase 12: Prepare the dissertation document (01/03/2023 - 30/06/2023
Condições
Research Conditions:
The work will be performed in the Instituto Pedro Nunes (IPNlis).
There is the possibility of awarding the internship with a scholarship, according to the candidate’s profile (around 875€ per month).
Observações
The work will be supervised by:
PhD. David Abreu (dabreu@ipn.pt), Instituto Pedro Nunes, Portugal<
The work will be co-supervised by:
Prof. Karima Velasquez (kcastro@dei.uc.pt), University of Coimbra, Portugal
Orientador
David Abreu
dabreu@ipn.pt 📩