Wednesday 19 March 2025
The quest for a more efficient and scalable way to embed virtual networks within physical ones has been an ongoing challenge in the field of network virtualization. Researchers have proposed various centralized algorithms, but these approaches often struggle with scalability issues as the size of the physical network grows.
A new decentralized approach, dubbed DeViNE, aims to address this problem by distributing the embedding process across multiple nodes within the physical network. This novel algorithm, described in a recent paper, has shown promising results in simulations and could potentially revolutionize the way virtual networks are deployed.
DeViNE operates by first selecting a random set of leaders from the physical nodes. These leaders then perform local embeddings, which involve finding the best possible mapping for each virtual node within their respective sub-networks. The algorithm uses a modified version of the ring-based election mechanism to ensure that all nodes agree on the optimal embedding solution.
One of the key benefits of DeViNE is its ability to handle large-scale physical networks with ease. By distributing the embedding process across multiple nodes, the algorithm can efficiently scale up to accommodate thousands of virtual nodes and links. This makes it an attractive solution for modern data centers and cloud infrastructure, where the need for high scalability is paramount.
Another advantage of DeViNE is its robustness in the face of node failures or changes in the physical network topology. Since each node is responsible for embedding a subset of virtual nodes, the algorithm can continue to function even if some nodes become unavailable. This level of fault tolerance is particularly important in production environments where downtime can be costly.
DeViNE also shows promise in terms of performance metrics such as acceptance ratio and revenue-to-cost ratio. In simulations, the algorithm outperformed centralized algorithms by up to 12% in acceptance ratio and achieved a revenue-to-cost ratio that was nearly 20% higher than other methods.
While DeViNE is still a relatively new approach, its potential implications for network virtualization are significant. By providing a decentralized solution that can scale to meet the needs of large-scale networks, DeViNE could help pave the way for more efficient and cost-effective deployment of virtual networks in the future. As researchers continue to refine and test this algorithm, it will be exciting to see how it evolves and whether it becomes a standard tool in the field of network virtualization.
Cite this article: “Decentralized Virtual Network Embedding with DeViNE”, The Science Archive, 2025.
Network Virtualization, Devine, Decentralized Algorithm, Scalability, Physical Network, Virtual Nodes, Fault Tolerance, Acceptance Ratio, Revenue-To-Cost Ratio, Cloud Infrastructure
Reference: Farzad Habibi, Juncheng Fang, “DeViNE: A Decentralized Virtual Network Embedding Algorithm” (2025).
