Can Neural Networks Provide Latent Embeddings for Telemetry-Aware Greedy Routing?
This work addresses the need for interpretable routing decisions in computer networks, offering a solution that balances efficacy with explainability.
The paper tackles the problem of explainability in telemetry-aware routing by proposing Placer, a novel algorithm that uses Message Passing Networks to generate latent node embeddings, enabling quick greedy routing and visualization of routing decisions.
Telemetry-Aware routing promises to increase efficacy and responsiveness to traffic surges in computer networks. Recent research leverages Machine Learning to deal with the complex dependency between network state and routing, but sacrifices explainability of routing decisions due to the black-box nature of the proposed neural routing modules. We propose \emph{Placer}, a novel algorithm using Message Passing Networks to transform network states into latent node embeddings. These embeddings facilitate quick greedy next-hop routing without directly solving the all-pairs shortest paths problem, and let us visualize how certain network events shape routing decisions.