TAGS: | | |

Say Yes To AIOps For Wi-Fi, Switching, And WAN

Sponsored Blog Posts

This guest blog post is by Trent Fierro, Senior Solutions Marketing Manager for AIOps at Aruba, a Hewlett Packard Enterprise company. We thank Aruba for being a sponsor.


Over the last few months, I’ve spent a lot of time speaking with IT networking teams on how their roles have shifted due to COVID-19. There’s almost nothing that hasn’t changed so far, yet a common theme persists – delivering the best user experience possible is still top-of-mind.

A paramount challenge, however, is solving problems remotely. For one, now that so many employees are off site, IT professionals have to rely on information from end users when troubleshooting – some of whom have very little technical expertise. For another, IT has had to deploy separate tools for managing their wireless, switching, and WAN infrastructure – each with their own dashboards, ways to view logs, and requirements for setting up service level expectations or thresholds.

Artificial intelligence is emerging as viable tool to help IT meet these challenges.

Artificial Intelligence – A New Paradigm

To simplify operations while keeping users happy, new solutions that include centralized cloud management, artificial intelligence (AI), and automation are rapidly gaining acceptance. Gartner refers to this shift as AIOps (Artificial Intelligence for IT Operations). In essence, it’s the use of insights from big data and machine learning to automate IT processes including event correlation, anomaly detection, and causality determination.

AI combines rich data, intelligence, and continuous learning to enable IT teams to proactively drive efficiencies through improved service delivery, operations, and network planning, to name just a few areas. AI is no longer just a vision; it is a critical and complementary tool that’s already in place in many facets of networks today. Whether your focus is remote work, supporting unified communications tools, or dealing with the explosion of IoT devices that require compute resources at the edge and cloud, AIOps can help.

Where To Start With AIOps

While AIOps solutions are available to monitor the behavior of your WAN, switching, and security deployments, assistance with wireless networks is traditionally where IT looks for help first. That’s because users will inherently blame the Wi-Fi network for poor connectivity, of which there are a number of things that can be the root cause, including channel interference, poor signal strength, latency, and too many users on the same access point.

Unless a problem is obvious, the traditional approach – which is no longer practical today due to social distancing precautions – would involve sending someone to conduct a site survey that reads not just signal strength, but examines other factors such as spectrum use, channel overlap, noise, and device placement. With this information in hand, a good wireless engineer can then sit with others to determine a plan of action. This plan might involve making a configuration change to see if the problem is resolved.

This is where an AIOps solution adds value. It uses data from all of the wireless devices and clients (user and IoT) connecting to the access points to dynamically create a baseline of network performance over time. Algorithms or machine learning models look for significant deviations in behavior. If a deviation is detected, advanced AI tools, such as Aruba’s AI Insights, automatically points out a problem, offers potential reasons for the issue, and recommends fixes. Where possible, problems are identified to help you fix an issue before it impacts productivity.

As an environment evolves to accommodate more (or fewer) people and devices, the system continuously analyzes the amount of traffic, the types of apps used, and any physical reconfiguration of the space, and then dynamically and automatically adjusts the baselines to ensure the network meets expected service levels. This replaces the need to manually reset thresholds per site, further minimizing false positives.

The Devil’s In The Data

In AIOps, data is the key to trusted insights and recommendations. In addition, the granularity of the data collected as well as the sources of data are important considerations. For example, a retail site will have a different usage profile than a convention center or a corporate campus. Your success will come from machine learning models that understand how your environment performs against similar profiles so that you have access to best practices that match your needs.

For AIOps to be effective, it’s also best that insights come from data that includes Wi-Fi, switching, WAN, and security to ensure that you have a holistic view. If user or IoT problems are being diagnosed, a wireless problem can result from an issue within a switch. For example, a wireless client with a DNS problem could stem from latency within the wireless or switching infrastructure. A holistic view that encompasses the wireless and wired domains will save time and avoid unnecessary guesswork.

Ultimately, AIOps and automation are about making an admin’s job easier and improving IT efficiency. It requires a continuous intake of data and baselining of your network to determine what is normal and what requires your attention. The ability to constantly learn from your network and compare data also makes it easy to identify intermittent issues that are difficult to resolve until they cause larger problems. By making it easier to monitor and diagnose your network automatically – regardless of whether you’re on-site or working remotely – the admins, users, and your organization will all benefit.

Aruba, a Hewlett Packard Enterprise company, can help you turn your data into actionable business value. To learn more, visit the link below or get in touch with us here.

https://www.arubanetworks.com/solutions/aiops/