Virtualize the NOC: Pass the Network Surge Stress Test with AIOps
Juan Perez | April 17, 2020

AIOps helps remote network operations teams collaborate effectively and adapt their architectures to changing traffic and usage patterns — during the current crisis and beyond.

AIOps helps remote network operations teams collaborate effectively and adapt their architectures to changing traffic and usage patterns — during the current crisis and beyond.

As the global pandemic radically alters how network services are consumed, NOC teams must quickly adjust to maintain the quality and reliability of business-critical applications. These new challenges stressing network architectures are impacting network managers both in large enterprises, as well as in service providers (NSPs), and vary in scope.

For example, some involve changes in workforce dynamics, such as the sudden need for NOC staffers to work from home, as well as staff fluctuations caused by the economic downturn. Others, of course, are technical in nature, such as the upending of network traffic and usage patterns caused by telework, which is causing unprecedented network management burdens.

BSS/OSS in a Digital World


Not to mention that this crisis found NOC teams already dealing with major challenges, as IT infrastructures became more dynamic, modular and complex due to digital transformation. This had made it increasingly difficult to monitor these IT environments, which in recent years had started to generate massive amounts of IT event data.

Thus, the existing complexity NOC teams are dealing with has become even more acute and made it harder for them to provide continuous service assurance.

“This complexity makes it extremely difficult to identify events which indicate an outage is about to occur,” said Moogsoft EMEA CTO Will Cappelli in “Pass the Network Surge Stress Test with AIOps,” the second webinar in his “Virtualize the NOC with AIOps” series.

The Global Network Infrastructure Under Stress


Amidst this “new normal,” with its short-term and long-term implications, AIOps has emerged as a lighthouse illuminating network architectures, and helping NOC teams work remotely and prevent service disruptions, said Cappelli, a global AIOps expert.

“I’ll share with you Moogsoft’s understanding of how the pandemic is impacting the tasks faced by network managers,” he said. 

“We’ll look at the difficulties coming in the wake of these changes, and outline how a solution based on AI technologies of a well-defined and specific sort can help,” Cappelli added.

Listen to the full webinar and get Cappelli’s comprehensive and detailed explanations of how Moogsoft AIOps can help network managers at large enterprises and NSPs create a “virtualized NOC” by:

  • Providing a full AIOps solution with the five key dimensions of AI for IT operations: data selection, pattern discovery, inference, collaboration and automation
  • Facilitating remote work, communication and data access across NOC teams with its unique Situation Room
  • Offering NOC teams a view of network services as a coherent whole with visibility into all layers of the stack

The webinar also included a Q&A with attendees. Here’s an edited transcript:


We’ve held back on AI-based solutions due to scale concerns. Is it true that AI technologies can’t deal with large volumes of data?

There’s a misperception that AI, machine learning and neural networks are all equal, and that neural networks take time to train, so consequently there must be a lot of latency in AI applications and they can’t handle large volumes of events. That’s just not true. 

Certainly neural networks are useful in some circumstances, and Moogsoft itself uses them for some causal analysis. But we also focus on nearly instantaneous pattern discovery. We also focus on ingesting events at scale and analyzing them very quickly. So yes, this has historically been a concern, but as AI has become more and more commercialized, particularly in IT operations, security and DevOps settings, we’ve seen those scale issues being dealt with.

You spoke about causal analysis, but root cause analysis has been a feature of OSS (Operations Support Systems) technologies for decades. What is different about what you are proposing?

You’re absolutely right. The big difference is that the classical approach to causal analysis was based on topology. You had a basic picture of your topology and you’d look at the events, and  and in which nodes they occurred at. If an event occurred at Node B and another at Node A, and if there was a meaningful connection between Node A and Node B, we’d say that there’s probably a causal relationship. 

Modern causal analysis doesn’t depend on topology. Topology is one of many elements used. It can also deal with very complex causal situations — multiple causes for a complex set of consequences. Modern causal analysis is based on being able to rapidly apply “what-if” analysis to see what changes. For example: If I modify this router here, what happens to my traffic volumes and to end user latency? The modern approach to causality is based on the maxim that there’s no causation without manipulation.

You have spoken about the urgency with which we need to deploy AI. This is still a big ask for a service provider or large enterprise. Is there some type of AI I can start with? What is the quickest path to value?

Go back to the five dimensions of AI. What we’ve found with our customers — even sophisticated enterprises and service providers — is that’s a lot to swallow at once. They have gotten the most value in their early deployments from two areas. 

First is the data selection process. Because there’s so much noise and redundancy in the data, there’s great value in getting down to those data items that say: “Hey, something’s taking place in this environment that you should be concerned with.” We’ve seen customers reduce that data volume in excess of 90%. 

The second one is the killer app in modern environments: the collaboration support. This is the ability to intelligently weave together a team, their tools and the information contained in those tools. That way, although they’re dispersed and must work in a virtualized NOC, the AI technology behind them weaves it all together. 

So start with data selection and collaboration, move on to pattern discovery, and then  to causal analysis. Once you’ve got those working smoothly, then you want to move on to automation.

Listen to the webinar

To get much more granular details about these topics and others, listen to a recording of Cappelli’s webinar.

Join us for part three of our “Virtualize The NOC with AIOps” webinar series, which is aimed at business leaders:

Futureproof Your IT Investment with AIOps on Wed, April 22 at 10 am PT

Moogsoft is a pioneer and leading provider of AIOps solutions that help IT teams work faster and smarter. With patented AI analyzing billions of events daily across the world’s most complex IT environments, the Moogsoft AIOps Platform helps the world’s top enterprises avoid outages, automate service assurance, and accelerate digital transformation initiatives.
See Related Posts by Topic:

About the author


Juan Perez

Longtime tech journalist turned digital marketer, Juan is now Moogsoft's lead content machine.

All Posts by Juan Perez

Moogsoft Resources

May 18, 2020

Applying AIOps to Logs Is Key for Observability

May 17, 2020

Moogsoft Enterprise 8.0: The Virtual NOC Is Here!

May 13, 2020

Rackspace Boosts IT Operations Management with AIOps

May 7, 2020

Assessing the Economic Value of AIOps