By Derek Handova

Artificial intelligence (AI) and machine learning will benefit network management and network monitoring going forward. And MSPs will need to know exactly how AI and machine learning will benefit them in order to use them effectively for customers.

In-house enterprise IT management will look more and more to outsource their network management and monitoring. And AI- and machine learning-powered analytics will be needed to manage and troubleshoot network issues as public cloud and private cloud models change.

Infinera’s Rob Shore

“To fully realize the vision of autonomous networking requires breaking free from the boundaries of traditional, closed hardware-centric architecture and moving toward more flexible solutions,” said Rob Shore, senior vice president of marketing at Infinera, supplier of intelligent network solutions. “From a cost perspective, AI and machine learning speed up resolution by minimizing downtime with preemptive maintenance allowing for putting proactive steps in place. Network monitoring reduces operational expenditure by reducing the number of steps and the need for skilled planning.”

In addition, AI and machine learning (ML) are in the process of revolutionizing network monitoring. Adding AI/ML to IT operations – including network monitoring – is commonly referred to as AIOps, a term coined by Gartner a few years ago.

Moogsoft’s Steve Kazan

“Given the growing scale and complexity of modern networks, AI/ML is becoming imperative for maintaining business agility,” said Steve Kazan, VP of channels and alliances at Moogsoft, a provider of AIOps solutions. “Intelligent automation of AIOps consolidates alerts into critical incidents, powers root-cause analysis, and helps ops teams take swifter action to remediate problems. AIOps delivers concrete business benefit, including lower mean-time-to-resolution, improved SLAs and cost reduction. For MSPs and MSSPs, AI/ML can be used to identify event patterns and trends across customers to resolve common issues. AI and ML are also being built into security products to identify anomalous behavior indicating malware, ransomware, breaches, and other network incidents.”

Security, AI, Machine Learning and Network Monitoring

When it comes to how AI and machine learning will benefit network management and network monitoring, MSPs and MSSPs should remember that no panaceas exist. Sophisticated human understanding must balance AI-driven automation.

Pondurance’s Landon Lewis

“AI and automation technologies extend a security analyst’s hands, allowing for quick analysis of massive amounts of data at scale — with efficiency no human analyst can match,” said Landon Lewis, CEO of Pondurance, a cybersecurity services provider specializing in managed detection and response. “Believing AI is the silver bullet that can address all cybersecurity challenges is as dangerous as the bad actors themselves. Although it can be used to detect unknown threats, AI still needs humans to provide reliable data to be effective. A lack of quality data leads to poor results. Even with quality data, trained AI tends to produce false-positives and is not good at explaining how it arrived at certain conclusions — it lacks the ability to understand context. For this reason, humans remain crucial.”

AI, Machine Learning, and Cause and Effect

Because AI and machine learning cannot determine causation – meaning that they are not able to tell why something happened – human understanding remains irreplaceable. And understanding “why” in outsourced managed cybersecurity services still has mission-critical written all over it, especially related to security incident investigations and analysis.

“Based on how AI is often marketed, many assume that AI-powered cybersecurity technology can simply replace humans,” said Jordan Mauriello, senior vice president at Critical Start, a security integrator, MDR and professional services provider. “And while its ability to ingest and process vast amounts of information is important, AI’s lack of causal reasoning is why human intelligence – especially from experienced security analysts – is still critical. Highly trained security analysts play an important role in …

… detecting, identifying and protecting against a range of cybersecurity threats — and will continue to do so for a long time.”

However, Mauriello goes on to state how AI and machine learning will benefit network management and network monitoring in important ways including the ability to correlate vast amounts of data from many sources — a must for informing security teams about incidents under investigation and making teams more knowledgeable and efficient at processing analytics.

“For example, AI/ML can inform you with anomalies, clustering and risk scoring of things to investigate,” Mauriello said. “This data can be used to better inform humans working to make decisions about security incidents. But AI/ML cannot make the decision for you.”

Filling Networking Monitoring Gaps with AI, Machine Learning

Modern networks are filled with mountains of information making it impossible to manually make heads or tails of most of it. But after learning basic patterns of IT traffic, AI and machine learning will benefit network management and network monitoring by cutting out the noise and exposing the signal for human decision makers — for everything from enterprise network bottlenecks to 4G and 5G service edge cases.

Verizon’s Matt Tegerdine

“In addition to monitoring and predicting network performance, AI and ML can help analyze variables such as signal strength when deploying new technologies, identify and fix equipment issues on devices trying to attach to the network, and help evaluate and draw conclusions for configurations of new cell sites,” said Matt Tegerdine, director of predictive analytics at Verizon. “All these examples require analysis of terabytes of data, and the patterns found using AI and ML on that data can add significant efficiencies.”

And it is vital for decision makers to distinguish between secure networks and safe networks when evaluating network security solutions. The industry has focused primarily on securing networks by preventing breaches: firewalls, passwords, two-factor authentication, phishing protection, endpoint security, and so on, according to Laurent Zimmerli, VP of product marketing at Open Systems, a provider of a secure SD-WAN. “AI and ML monitoring tools may help identify breaches before they happen, but no network will ever be 100 percent secure,” Zimmerli said.

Overall, MSPs and the customer networks they operate are exposed to a wealth of operational and network data and yet only leverage a part of it. This is where analytics, automation, and AI kicks in.

Ericsson’s Jonas Åkeson

“Combined with telecom domain expertise, we will now be able to make use of, learn from and benefit as a result of all the data networks that we operate,” said Jonas Åkeson, head of automation and AI, managed services at Ericsson. “As a result, operations and services will become fully data-driven. And data-driven will give us truly predictive, proactive and preventive operations, moving us from cost-centric to fast and value-centric.”

And because AI continuously leverages the most recent data points to make predictions, new patterns in data emerge constantly and must be taken into consideration when planning for future network needs.

“Traditional methods may take several weeks or months to identify a new pattern in data,” said Greg Cox, CTO architect at Sungard Availability Services. “AI can recognize the new pattern by the second or third recurrence in the same data flow and adjust future predictions accordingly.”