Greg Mayo

Open the pod bay doors, Hal.”  “I’m sorry, Dave. I’m afraid I can’t do that.” 

That infamous movie dialogue, from the classic 1968 film, 2001: A Space Odyssey by director Stanley Kubrick, depicts the fictional exchange between stranded astronaut Dave Bowman and Hal, the artificial intelligence (AI) powered super-computer controlling his spaceship.

This human-like intelligence is the public’s perception of AI, but the reality is not nearly as cool yet, as it continues to develop into a mature technology with many applications.

AI basically makes machines more intelligent. Examples include cognitive computing, visual recognition, natural language processing, etc. When it comes to network service assurance and security, AI plays an important part. At the same time, machine learning (ML), which is a subset of AI, is responsible for unsupervised functions that use algorithms to learn from how data is related to other aspects of data. ML can also be supervised, involving either simple neural networks or complex neural networks. In these cases, a program is trained to recognize the difference between good datasets and bad, then is enabled to learn from new datasets it encounters.

This type of intelligence is highly effective for specific use cases, such as attenuating spectral frequencies in the RAN or identifying very detailed pattern anomalies, but less so when dealing with large-scale complex problem solving found in communication service provider’s cloudified network environments with 5G. 

Data Explosion Ends Up with Finger Pointing in the War Room

What makes today’s communications service provider’s (CSPs) network environments so challenging is the explosion of data. With the high volume of flows across networks - down to the packet, including individual sessions, IoT flows, etc. - there's so much data; if Smart Data is not applied, AI/ML algorithms will potentially take data lake teams down paths that aren’t effective. This results in an inefficient use of time across departments leads to finger-pointing in the war room during the isolation process and ultimately can end up taking days to resolve problems. 

Using AI/ML as a Tool for Automation

Due to the complexity of CSP’s networks and services, there is a tremendous need for intelligent automation sequences to improve war room efficiency. By automating the gathering of information and running analytics to glean the most pertinent information, the root cause of issues can be isolated, finger-pointing can be mitigated, and problems solved more quickly. 
Intelligent automation sequences blend AI/ML techniques, combining ML libraries for classification and outlier detection and knowledge base and inference engines to conduct deep correlation and case-based reasoning. The objective is to provide an automated consultant that can offer concise findings to the war room teams. 

Leveraging Smart Data to Produce Smart Outcomes

NETSCOUT’s Omnis Automation platform is designed to solve complex problems using AI and ML, delivering actionable business intelligence. This automated solution takes raw traffic data and refines it into a clean data set we call Smart Data. Omnis Automation then applies AI/ML algorithms with built-in domain expertise that has a deep understanding of the CSP’s network. 

Data is subsequently run through microservice-based analytics chains to detect the most relevant outliers. The use of NETSCOUT’S extensive domain knowledge produces smart outcomes. In short, Omnis Automation automates tedious tasks, freeing up teams in the war room to focus on assuring service and achieving security objectives.  

Subscribe to Our Blog