Government IT agency wants AI to predict network disruptions

The Malta Information Technology Agency (MITA) is turning to Artificial Intelligence in a bid to automate its problem-solving capabilities for network disruptions, as well as predict surges in traffic bandwidth traffic.

MITA provides information technology services to the Maltese government and assists the State in technological innovations.

The agency now is tapping the private market to find out which AI-powered tools can be integrated into its network management system, the critical 24x7 monitoring service for the government’s IT systems.

MITA’s infrastructure comprises around 2,500 servers, 4,500 network devices, 600 websites and handles approximately 19 terabytes of network traffic daily.

Now it wants to know how AI can integrate with the various tools that are used to monitor the performance and workloads of its servers and cloud infrastructures.

Specific case scenarios the AI tools could address would be include the predictive generation of alerts when there are abnormal spikes in network traffic during off-peak hours. While monitoring tools can show the sudden increase through a spike, they have no pre-configured thresholds to be able to detect the issue or generate alerts.

In other typical examples where AI tools might assist MITA would be in cases of intermittent network outages, which might last for just three minutes. Despite the significant impact this could have for users working in 24x7 offices, there might not be enough time for NOC to manually analyse the incident.

AI can also hep in generating alerts related to critical system failures and service disruptions across multiple servers, allowing the NOC to identify and address incidents in a timely manner, minimising service disruptions and potential business impact.The enhancements will allow MITA to have an analysis of all the observability data, with automated predictions, advanced anomaly detection and trend analysis, all in a bid to reduce the likelihood of incident occurrence and resolution times.