DNFS-VNE: Deep Neuro Fuzzy System Driven Virtual Network Embedding

Overview

• Network virtualization is a promising solution for meeting diverse demands and ensuring differentiated quality of service (QoS)
• Virtual network embedding (VNE) is a critical enabling technology for network virtualization
• Existing deep neural network (DNN)-based approaches for VNE have limitations in terms of interpretability and transparency
• Interpretable deep learning represented by deep neuro fuzzy systems (DNFS) can help address these limitations

Plain English Explanation

Network virtualization is a way of decoupling the physical resources that make up a network, like servers and routers, from the services and applications that run on that network. This allows the network to be more flexible and adaptable, and can help ensure that different users or applications get the level of service they need.

Virtual network embedding (VNE) is a key technology that enables this flexibility. VNE is about figuring out how to assign the physical network resources to the different virtual services and applications in the best way. However, the existing DNN-based approaches for VNE have some limitations. They tend to be "black boxes" - it's hard to understand how they are making decisions.

This is a problem, especially in industrial settings like the Industrial Internet of Things (IIoT), where it's important to be able to interpret and explain the decisions being made. A new approach called interpretable deep learning, represented by deep neuro fuzzy systems (DNFS), may help address this by providing more transparency.

Technical Explanation

The paper proposes a DNFS-based VNE algorithm that aims to provide an interpretable network virtualization scheme. The key elements are:

• Data-Driven Convolutional Neural Networks (CNNs): These are used as fuzzy implication operators to compute the embedding probabilities of candidate substrate nodes through entailment operations. The identified fuzzy rule patterns are then cached into the CNN weights through forward computation and backpropagation.

• Fuzzy Rule Base: The fuzzy rule base is constructed based on Mamdani-type linguistic rules using descriptive linguistic labels, rather than numerical values.

• DNFS-Driven Policy Network: A five-block structure-based policy network serves as the agent for deep reinforcement learning (DRL), which optimizes the VNE decision-making through interaction with the environment.

The effectiveness of the evaluation indicators and fuzzy rules is verified through simulation experiments.

Critical Analysis

The paper presents a promising approach to address the interpretability limitations of existing DNN-based VNE methods. The use of DNFS provides more transparency into the decision-making process, which is important for real-world applications like IIoT.

However, the paper does not provide a detailed comparison to other interpretable DL approaches, such as attention mechanisms or feature importance analysis. It would be helpful to understand the relative strengths and weaknesses of the DNFS-based approach.

Additionally, the paper focuses on the algorithmic aspects and does not delve into practical deployment considerations, such as the computational overhead of the DNFS-based model or its scalability to large-scale networks. These are important factors to consider for real-world adoption.

Conclusion

This research proposes an interpretable deep learning-based approach for virtual network embedding, which addresses a key limitation of existing DNN-based methods. By leveraging DNFS, the proposed algorithm provides more transparency into the decision-making process, potentially making it more suitable for industrial applications like IIoT.

The technical details and experimental validation suggest that this is a promising direction for enhancing the flexibility and explainability of network virtualization. Further research is needed to benchmark the approach against other interpretable DL techniques and to explore the practical deployment considerations.