Automation technology has long been a part of IT’s arsenal, as enterprises work to simplify increasingly complex networking and operational environments. Looking ahead, the more AI is added to the automation mix, the more enterprise operations teams stand to gain.
“2025 will be the year for AI initiatives, where AI-powered automation has reached a tipping point from being a nice-to-have to a requirement,” said Bill Lobig, vice president, product management, with IBM Automation.
“Automation is needed to solve AI’s complexity. Organizations can now confidently advance and scale their AI initiatives using automation, moving from spending time managing and maintaining AI applications and IT environments, to proactively detecting and resolving issues. Automating these tasks will be critical for competitive advantage,” Lobig said.
“Next year, you won’t be able to have an AI conversation without talking about automation, and vice versa,” Lobig said.
[Read more: 5 network automation startups to watch]
Customers will be able to apply AI to data to see patterns and trends, and AI will improve the signal-to-noise ratio as companies try to sift through the metrics, logs, events, and traces that are spewing out of customer systems, Lobig said. “Most of these systems are now more complicated than ever, with hybrid cloud becoming the way many organizations operate,” Lobig said.
AI-driven automation could help close staffing gaps
Research firm Enterprise Management Associates (EMA), too, cites the infrastructure complexity that has resulted from hybrid and multi-cloud networks and the need for more advanced automation.
“EMA research finds that hybrid clouds are particularly problematic for network operations teams today. They’re struggling with complexity there. Better automation could help,” said Shamus McGillicuddy, vice president of research at EMA.
Today’s vendors are developing AI and ML algorithms that can detect anomalies, identify the root causes of problems, and predict changes in network utilization, McGillicuddy said.
“By integrating these AI insights with network automation tools, IT organizations can trigger automated workflows that resolve problems, adjust capacity, and make other proactive changes to ensure infrastructure resilience.”
That’s more important than ever, especially given that IT organizations are often understaffed, particularly in network engineering, McGillicuddy said. “They’re also tasked with supporting new initiatives and reducing costs. AI-driven automation can help them close staffing gaps, reduce costs, and be more responsive to changing business need,” McGillicuddy said.
There are dozens of automation tools available today – including Yang, Ansible, Terraform and Puppet, to name just a few – and vendors are in different stages of development in terms of AI-enabled capabilities.
“There are many companies leveraging AI-driven automation today. However, vendors vary in their level of maturity with AI today,” McGillicuddy said. “Some are just getting started. Others have been delivering solutions for a couple years. EMA research shows that only 42% of organizations trust the AI-driven capabilities of their network observability tools today.”
Network teams are happy to leverage AI insights, but they’re not ready to remove humans from the equation. Closed-loop, AI-driven automation will not be mainstream in 2025, McGillicuddy added.
AI-driven tools will slash network configuration errors, Cisco predicts
More than 40% of network outages are caused by misconfigurations, and those outages can cost businesses 9% of their annual revenue, stated Liz Centoni, Cisco’s executive vice president and chief customer experience officer.
“As such, one of the most promising developments on the horizon is the potential for AI to virtually eliminate these manual misconfiguration mishaps,” Centoni wrote in a blog about 2025 technology trends.
“Intelligent, automated tools can execute automated workflows throughout the network lifecycle and provide traceability for every action. AI-driven tools are set to revolutionize network management and assurance, learning and benchmarking from each configuration to reduce errors and ensure uninterrupted operations,” Centoni wrote. “We will see misconfigurations decline rapidly as AI adoption grows, making automation accessible to more organizations, and we expect to see network downtime caused by human error rapidly approach zero.”
Just as vendors are at different stages of AI-driven automation, so, too, are enterprises.
The ultimate impact of AI and automation really depends on where customers are in their digital journey, with some still managing networks manually and others more advanced, said Matt Gillies, Cisco Fellow, vice president, and chief architect, global solutions.
“Any time there’s a manual operation, it’s an opportunity to automate,” Gillies said. Hyperscalers such as Microsoft and Meta have built their own automation tools, “and they’ve done that for a long period of time because of the velocity of that business – they need to scale rapidly,” Gillies said. “But we have a lot of customers that still manage their networks with spreadsheets and swivel chairs.”
“But the cost of outages related to misconfiguration, in particular, has just become too risky for many businesses, and so I would say there’s a renewed focus on automation,” Gillies said.
Network observability boosted by AI
Another area where automation can potentially make a big impact is network observability.
Cisco ThousandEyes provides digital experience monitoring to give customers greater insight into their network operations. ThousandEyes network intelligence agents are software probes placed throughout the enterprise network infrastructure, including data center endpoints, routers and switches, cloud-based resources and branches.
Cisco says there are tens of thousands of ThousandEyes agents spread across the internet and enterprise networks, and the platform is powered by more than 650 billion daily measurements globally.
“Being able to automate network performance management and respond to trouble spots quickly will help organizations solve problems quickly, so customers won’t be waiting for help desk calls. [The] pace of response will be so much faster than it ever was before,” Gillies said.
Adding AI to the ThousandEyes package will improve the software’s ability to spot the source of a problem and prioritize which incidents require immediate attention, Gillies added.
A need for greater network and system visibility is also an important driver for enterprises to consolidate some of their myriad management tools.
“Many NetOps teams will replace and/or consolidate the tools they use to monitor, troubleshoot, and optimize their networks,” said EMA’s McGillicuddy.
“They’ll look for solutions that can collect large and diverse data set and apply AI and advanced analytics to deliver advanced event correlation, intelligent alerting, and automated root cause analysis. Much of this consolidation will occur in the network management tools, but many IT organizations will try to consolidate across IT teams to enable a cross-domain observability portfolio,” McGillicuddy said.
AI, automation and high-performance workloads
AI-driven automation gains could also be directed at making AI workloads easier to deploy, run and manage, IBM’s Lobig said.
One example is in Kubernetes environments, which are becoming the default runtime choice for AI workloads and LLMs.
“Kubernetes automatically scales up and down, but it does so by spawning and or destroying nodes. And nodes have a fixed, defined configuration [with] memory, storage, compute and GPUs attached. And so, customers could have a situation where they need more memory, spawn up a new node, and that new node definition has like eight GPUs attached to it, because that’s a statically defined thing. And suddenly they are paying for all these GPUs that they are not using,” Lobig said.
“AI and automation can bring intelligence to how workloads are scheduled and scaled across large environments,” Lobig said.
Source:: Network World