An Intelligent Data-Plane with a Quantized ML Model for Traffic Management
Bibliographic record
Abstract
Offloading some of the traffic management decision-making functionalities to intelligent data-planes (IDPs) can significantly enhance the accuracy and adaptation speed of network services. An IDP executes, at line-speed, one or more machine learning (ML) models for real-time inference and decision making. Unfortunately, existing IDP deployments either realize only a limited set of ML models such as decision trees or require substantial modifications in the switch hardware. These limitations can be attributed to the inherent scarcity of both the computational and memory resources and the strict high-speed per-packet processing demands. To address the aforementioned limitations, we propose a novel ML-based management framework, the in-network quantized ML architecture (INQ-MLA). First, INQ-MLA delegates the task of training and continuously optimizing the IDP ML model to the control-plane. The latter adopts a tailored quantization-aware training process to compensate for the effect of precision loss due to quantization. Second, INQ-MLA employs an efficient quantization mechanism to transform the trained ML model parameters (e.g., weights and activation functions outputs) from floating-point representations to smaller low precision fixed integer values that can be easily processed and stored in the data-plane. Finally, INQ-MLA ensures that the deployed ML model is integrated into the IDP pipeline by limiting all its execution operations to simplified arithmetic operations that are available in most switches. We developed a proof-of-concept implementation of our proposed architecture using P4-based switches. Experimental results demonstrate that INQ-MLA can achieve a high-level of accuracy at runtime.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".