If you’re an IT professional, you’ve heard of AIOps or artificial intelligence for IT operations. The term, originally devised by Gartner in 2017, describes the process of managing data from an application environment using artificial intelligence. While the concept is simple, AIOps is actually quite difficult to put into practice.
AIOps combines machine learning (ML), behavioral ethics, and predictive analytics, and it involves massive amounts of telemetry information being generated by network devices. There are many automation tools out there that work well in static environments. But modern environments aren’t static and are constantly changing, which means organizations need tools that can keep up with the changes.
A new 2021 State of AIOps Study, published by ZK Research in partnership with Masergy, found IT departments spend nearly 50 percent of their time monitoring app performance and network troubleshooting. It takes IT departments on average 30 minutes to fix and resolve customer-facing issues. That’s a half hour of interrupted service.
With AIOps, critical application performance is significantly improved. The critical apps are always observed by the ML system. The traffic patterns are understood, and the network responds to those needs dynamically.
I recently spoke with Ajay Pandya, director of product management at Masergy, about the study’s key finding and how AIOps trends are paving the way for fully autonomous networks. Highlights of my ZKast interview, done in conjunction with eWEEK eSPEAKS, are below.
- 2021 State of AIOps Study was conducted to understand where AIOps has business value for IT and where the technology is headed in the future. ZK Research surveyed 500 U.S.-based IT decision makers and C level executives across seven verticals: technology, manufacturing, retail, healthcare, finance, media and communications, and pro services. The study focused on enterprise-class companies with $250 mil. to $10 bil. in revenue.
- A few surprising findings: 64 percent of IT leaders said they’re already using AIOps, while 55 percent said they’re using AIOps for both networking and security. The majority (more than 90 percent) believe AIOps plays an important role in managing the network.
- The study also found organizations are benefiting from AIOps. The top use cases for respondents include cloud app analytics, performance improvement, network service optimization, and faster threat detection and response. Sixty-four percent of the respondents measure AIOps success based on IT operational efficiencies, and another 54 percent on improved network or app performance. This shows there is real business value in an AIOps investment, both making AI more efficient and improving business operations.
- Additionally, 84 percent of IT leaders see AIOps as a path to fully autonomous networks. In fact, 86 percent expect to have a fully automated network in the next five years and 97 percent are confident that AIOps can be trusted to act alone. These predictions may not happen in the next five years, yet AIOps will be a major investment area.
Key Takeaways from the AIOps Study
- The era of AIOps is here. Companies that haven’t started looking into it yet need to start the process now, before they’re left behind. Although AIOps can lower costs, it must be implemented for the right reasons, such as IT transformation.
- While IT leaders recognize there is a problem, the findings don’t reflect true use of AIOps. Many solutions claim to be AIOps but are actually rules-based systems. AIOps is a closed loop system, which takes incorrect data and feeds it back into the system to become part of the training set. This is what makes AIOps smarter over time. IT buyers need to be careful and make sure they’re choosing the right solution.
- AIOps requires both software-defined networking (SD-WAN) and secure access service edge (SASE). Companies should find providers that bring those together in a central place for automation. Most organizations trust AIOps to run their networks. However, organizations won’t be fully autonomous until they move their network and security into software, which requires both SD-WAN and SASE.
- There is a strong link between AIOps and SD-WAN. A non-SD-WAN network is not centrally managed or centrally orchestrated since data is siloed. SD-WAN provides the centralization needed to make networks more efficient. In the study, more than two thirds of companies (73 percent) identified SD-WAN modernization as a top investment and as a prerequisite for AIOps.
- Next-gen enterprise networks are perfect candidates for AIOps, whereas legacy networks weren’t designed for it. SASE will have a larger presence in enterprise networks in the next two to three years. That’s where AIOps will come into play. According to 70 percent of IT leaders, AIOps performs really well in a SASE architecture.
- AIOps is a journey rather than a destination. As apps move to the cloud, a digital transformation will take place within networks. Organizations are headed toward self-healing, self-managed networks and IT leaders are putting more trust into AIOps to help with day-to-day operations.