Network Autonomous deployment using artificial intelligence
At Discovery – Future Network Labs, we are at the forefront of
autonomous network deployment, working on use cases that enable the automatic and efficient deployment of new network elements. To tackle these challenges, we implement deployments using Kubernetes, a
platform that allows for the automated management, scaling, and maintenance of containerized applications.
By containerizing network functions, we can continuously monitor the load and memory levels of network elements, regardless of the Core provider, and anticipate potential critical scenarios through predictive analysis. Our models allow us to precisely identify when and where an overload will occur.
The main steps to achieve a closed-loop automation flow can be summarized as follows:
1) Network Monitoring: By sending events, we monitor key network metrics and external factors to generate datasets once the data is processed and transformed.
2) Data AI Analysis: We generate and manage AI models based on the collected network data. It is important for the MLOps system to be consistent so that the trained models successfully adapt to each scenario and profile, considering the variability of each situation.
3) GitOps Configuration: We rely on the GitOps methodology for application and infrastructure management, using Git as the configuration source for automation and environment management, achieving automated system operation.
4) New Network Deployment: A new UPF with the selected configuration will be deployed to resolve congestion, and monitoring will continue to identify when it is no longer necessary. We can also apply the best configuration according to the new scenario.
In conclusion, the ability of networks to dynamically adapt to changing demands is essential for maintaining service quality and continuity. AI and automation allow us to anticipate resource issues and optimize their usage. At Discovery – Future Network Labs, we are researching innovations to improve operational efficiency, paving the way for the networks of the future and the present.
By containerizing network functions, we can continuously monitor the load and memory levels of network elements, regardless of the Core provider, and anticipate potential critical scenarios through predictive analysis. Our models allow us to precisely identify when and where an overload will occur.
The main steps to achieve a closed-loop automation flow can be summarized as follows:
1) Network Monitoring: By sending events, we monitor key network metrics and external factors to generate datasets once the data is processed and transformed.
2) Data AI Analysis: We generate and manage AI models based on the collected network data. It is important for the MLOps system to be consistent so that the trained models successfully adapt to each scenario and profile, considering the variability of each situation.
3) GitOps Configuration: We rely on the GitOps methodology for application and infrastructure management, using Git as the configuration source for automation and environment management, achieving automated system operation.
4) New Network Deployment: A new UPF with the selected configuration will be deployed to resolve congestion, and monitoring will continue to identify when it is no longer necessary. We can also apply the best configuration according to the new scenario.
In conclusion, the ability of networks to dynamically adapt to changing demands is essential for maintaining service quality and continuity. AI and automation allow us to anticipate resource issues and optimize their usage. At Discovery – Future Network Labs, we are researching innovations to improve operational efficiency, paving the way for the networks of the future and the present.
Autonomous network deployment flow using IA
