Toward An Experimental Federated 6G Testbed: A Federated Leaning Approach

Kholidy, Hisham A.; Hariri, Salim

doi:10.1109/aiccsa56895.2022.10017506

Cited by 7 publications

(2 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this use case, we use 8 KPIs criteria/attributes for 20 EPoPs candidates as depicted in Fig. 2 and are computed in our 5G security testbed [8] as described below.…”

Section: B the Weighted Hierarchical Structurementioning

confidence: 99%

“…we develop a Federated Network Slicing Orchestrator (FNSO) that uses the HF-TOPSIS to rank and select the proper EPoP that can host the network slice using numeric and fuzzy KPIs such as the security level, cost, and performance criteria that are computed and weighted using a newly developed weighted hierarchical structure of the KPIs described at Section 4.b. The accuracy and performance of the proposed FNSO and HF-TOPSIS are tested and evaluated using our new federated 5G security testbed [8].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Dynamic Network Slicing Orchestration in Open 5G Networks using Multi-Criteria Decision Making and Secure Federated Learning Techniques

Kholidy

2024

Preprint

View full text Add to dashboard Cite

Network slicing enables new revenue opportunities for service providers by allowing them to offer customized network services tailored to various industry verticals and use cases. 5G network uses slicing technology to enable the creation of tailored network slices to support diverse use cases, such as enhanced mobile broadband (eMBB), massive machine-type communications (mMTC), and ultra-reliable low-latency communications (URLLC). However, the complexity of network slicing also introduces challenges, such as maintaining service-level agreements (SLAs), quality of service (QoS), security across multiple slices, and service provisioning and dynamic allocation of resources. This paper develops a novel Federated Network Slicing Orchestrator (FNSO) that uses the Multi-Criteria Decision Making (MCDM) method to rank and select the proper telecommunication service provider that runs at certain edge Points of Presence (EPoP) of the 5G-based IOT network for slice deployment. The proposed FNSO integrates the Hexagonal Fuzzy approach with the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) to address the shortcoming of the TOPSIS and to rank and select the proper EPoP that can host the network slice using numeric and fuzzy Key Performance Indicators (KPIs) such as the security, cost, and performance criteria. Furthermore, we augment the FNSO with a Secure Federated Learning (SFL) model to protect the local 5G domain service provisioning KPIs and secure the FNSO against potential attacks. The experiment results depict that the average slice acceptance ratio of the FNSO is higher than the current solutions such as GRU-DNN, VIKOR-CNSP, and T-S3RA by 28.2%, 10.13%, and 6.1%. Furthermore, On average, the SFL model is faster than the DeTrust-FL, HybridAlpha-FL, PHE-FL, and HybridOne-FL by 7.11%, 7.21%, 15.73%, and 20.61% respectively, and slower than Classic-FL, on average, by 4.37% due to the secure aggregation that the SFL offers.

show abstract

“…In this use case, we use 8 KPIs criteria/attributes for 20 EPoPs candidates as depicted in Fig. 2 and are computed in our 5G security testbed [8] as described below.…”

Section: B the Weighted Hierarchical Structurementioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%