The why and how of AI-enhanced infrastructure – Cisco Blogs


Part 2 of the Series: AI-Enhanced Infrastructure for AI Ops.

In this blog, I’ll continue exploring how an AI-enhanced infrastructure can be used by IT operations teams today.  I’ll discuss the technologies that are involved and how they are powering an AIOps model for ITOps and NetOps teams.

What are AI-enhanced technologies?

I would like to establish a shared definition: When we talk about technologies powered by AI, we are referring to a set of tools and systems that allow IT or network operators to drastically simplify their activities. This is especially relevant where analysis of the available data is highly complex.

Think about the threshold you establish in a traditional, day-to-day IT and network operations workflow to detect abnormal behavior. For example, the onboarding time for a new mobile device on your Wi-Fi network, which your organization has defined as up to say, 5 seconds.

The reality is that the 5 seconds number is a static threshold that does not have a basis in empirical data. A data-driven dynamic threshold might show that onboarding could actually be done in just 2 seconds. Or that the onboarding time will be different if the user is in a location with 5 connected clients, compared to another location with 5,000 clients, by simply factoring the different demands that the infrastructure would have in those two scenarios.

In this example and in any case where ITOps team needs to analyze high-volume data, that is when an AI-enhanced system can help your organization make better, faster decisions. It aids human operators with a better understanding of the events happening within the network in a simple and actionable way. It helps with remediating potential failures and errors even before they occur. I’m really excited about the progress we are making with AI-enhanced technologies that make a real difference for network practitioners.

There are multiple categories of AI-related technologies that can be applied to AIOps models, but we can commonly consider two foundational categories for any AI-enhanced infrastructure model:

  • Machine Learning (ML): Machine learning is a data analysis method that automates the construction of analytical models. It is a branch of AI based on the idea that systems can learn from data, identify patterns, recognize abnormal behavior, and make decisions with minimal human intervention.
  • Machine Reasoning (MR): A category of technologies derived from expert knowledge that allows a computer to work through complex processes that would normally require a human. Common applications for MR are time-consuming and tedious detail-based workflows that require selecting the best scenario based on the many options available.

As I have mentioned before, the accelerating scale of data coming from your network and the multicloud environments in which your organization operates have the potential to provide insights into the health of your infrastructure. When this is combined with the growing expectations of users for a seamless IT experience, ITOps teams are being forced to extend the human capabilities of their staff by leaning more and more on machine-driven skills and processes.

And one of the areas where the value of these systems is clear is in freeing up network practitioners to focus on the development of new services. The machines will help teams off-load repetitive tasks, and focus on the architecture design for the new services that depend on human IT talent, experience, and creativity.

The true potential of AI-enhanced technologies is in the right combination of IT staff supported by a ‘trusted’ AI system.

An AIOps infrastructure model must build trust by allowing ITOps and NetOps personnel to understand each step in the analysis and decisions delivered by AI. For it to be useful, the system must be able to deliver the analysis in the most accessible terms for human understanding. This is the virtuous cycle that will allow the capabilities to grow from a traditional infrastructure to a truly AI-powered model.

If you are interested in hearing more about how AI technologies are supporting IT teams today, don’t forget to register and join us at the upcoming event below. Looking forward to seeing you there!

Join our experts to discuss tried and tested AI-enabled capabilities
that save time and headaches for many organizations today.
Register for the webinar:

January 25th | 10AM Pacific Time

 

Additional resources

Share:



Source link