Industry: Telecommunications | Year: 2018
The Trento team of Spindox Labs brings artificial intelligence and big data to TIM’s DI-MON (Digital Monitoring). Thus a tool created to observe in real time the use of hardware resources and the emission of application events generated by business processes becomes predictive. The data is in fact aggregated to allow for calculations based on specific systems behavior patterns.
TIM’s artificial intelligence module for digital monitoring has been designed to communicate directly with the customer’s log management stack. The latter was developed in turn by Spindox. The predictive analytics engine uses machine learning and deep learning techniques and offers three features:
- Anomaly Detection: learns the trend of the metrics measured in normal situations and detects any deviations, indicating their degree of criticality.
- Correlation: suggests associations between anomalous events that occur on different metrics.
- Prediction: estimated probability of anomalies in a short future time interval.
Being a machine learning system, the engine is able to refine its predictive models over time. In other words, the performance and capabilities of the machine improve over time.
The main purpose of the artificial intelligence module is to provide automatic tools to understand anomalous events that can occur on all systems under observation.
In particular, the engine offers the following advantages:
- Reduction of the time margins necessary to make important decisions and therefore increase the efficiency of the operating machine (realtime decision making).
- Reduction of the effort required for application monitoring, with further advantages from the point of view of efficiency.
- Immediate reaction to any problems that arise in the business process.
- Prevention of the problems themselves, through immediate recognition of changes and early analysis of anomalies.
- Simpler definition of standard operating protocols