AI and Machine Learning for Networks

AI and ML methods apply to many network and data center problems, from querying network state and validating configurations to forecasting traffic and classifying attacks.

This ebook by Tomasz Janaszka maps each method class, from generative AI and time series forecasting to clustering and reinforcement learning, to the data it works on and the problem it solves, with popular algorithms and the challenges to plan for.

The main topics covered in this e-book:

GenAI assistants for network state queries, configuration validation and troubleshooting
Time series forecasting methods, from ARIMA and Prophet to Temporal Fusion Transformer
Regression models linking QoS metrics such as jitter and packet loss to QoE
Clustering and anomaly detection over trouble tickets and device configurations
Reinforcement learning to optimize OSPF and IS-IS link weights

AI and Machine Learning for Networks — document render

Why download this e-book?

Which AI and ML methods suit network data?
The ebook maps generative AI, time series forecasting, classification, regression, clustering, anomaly detection and reinforcement learning to the network data and problems each one handles.
Where can generative AI help network engineers?
GenAI assistants and chatbots can support network troubleshooting, anomaly detection, configuration management and reporting. The ebook lists the GenAI tools and frameworks worth your attention.
Which algorithms are worth trying?
For each class of problem the ebook lists popular algorithms and frameworks, such as reinforcement learning to optimize routing metrics in networks running OSPF or IS-IS.
What should you plan for before you start?
The ebook names the challenges to prepare for: data quality and imbalance, feature extraction, model interpretability, scalability, real-time processing and the resources training needs. It also explains why data scientists, ML engineers and domain experts need to work closely together.