Chris MacFarland, Chairman & CEO, Masergy, a Comcast Business Company.
Whether training for a 5K, practicing yoga or studying academic subjects, learning takes discipline. Knowledge and proficiency always come with practice, and the same holds true for artificial intelligence (AI) and machine learning (ML). The algorithms get exponentially smarter with practice; over time, they get better at predicting and generating optimal business outcomes.
This is one reason I’ve been a strong proponent of starting early with AI for IT operations (AIOps) technologies. With hybrid work becoming the new norm, AIOps becomes imperative. Training is a primary step on the path to automation.
The Tool For Secure Hybrid Work
Hybrid work has approached the point of no return. The edges of corporate networks are exploding, and IT organizations are hyper-challenged to ensure business continuity and data security for remote workforces. Gartner predicts that by the end of 2021, more than half of all knowledge workers will be remote. Delivering globally reliable application performance and protecting data everywhere has become IT’s priority job, and it’s moved beyond the point of human acumen.
That’s why, in 2020, the AIOps market was valued at $13.51 billion in 2020 but is projected to reach $40.91 billion by 2026, according to Mordor Intelligence. With data flowing through a mesh of networks and cloud services to remote workers, it’s nearly impossible to find and tackle IT issues without assistance from AI. Even with unlimited budgets, the answer should not be “throw more people” at the problem — the human brain is unfit for the job. The task at hand is disproportionate to our collective ability to effectively manage this undertaking.
Path To The Intelligent Future
Imagine an IT infrastructure so smart it could foretell the future by observing how all applications are performing and how all users are behaving. With AIOps, we can leverage real-time information to benefit the network and security. As I mentioned in a previous article, the power of AIOps is being unlocked, particularly as it converges with the network and security in a new category of SASE solutions. One study sponsored by my company reveals the IT community is confident in taking this exciting ride into the future with 97% of respondents trusting AIOps tools to act on their own recommendations — the key to full autonomy.
What’s the impact? Almost daily news related to security breaches or cyber surveillance provide stark reminders of the corporate duty to secure hybrid work. Security arguments center around the cloud and endpoints, but that is only the tip of the iceberg. Underlying networks and user trust also play an active role in security equations. According to IBM research, most incidents are contained only after 287 days, and the average breach cost is $4.24 million in 2021 — the highest in 17 years.
I’m most excited by the opportunity for AIOps to identify and remediate anomalous behavior, because networks are the keeper of secrets, in the form of forensic files. No matter where the workload or application resides — the cloud, on-premise or a hybrid mix — AIOps can learn from traffic patterns to trigger faster threat response and improve performance. This intelligence is available today, including recommendations for improvement. Now, it’s time to prepare for AIOps.
Training For Automated IT Operations
Transparency, centralization and agility are the muscles to condition for the long run, because overarching intelligence and control are fundamental in making IT systems smart enough to recognize security attacks and reroute applications for automated business continuity.
A powerful combination of AIOps, networks built on software (not hardware) and security capabilities are needed to activate our hybrid work today and into the future. Seventy-three percent of IT leaders identified software-defined network modernization (SD-WAN) as the top IT investment required to prepare for AIOps in the same study sponsored by the company mentioned above. Why? AIOps doesn’t work in legacy networks. It’s blinded by silos and segmented environments.
• AIOps is only as smart as the data it is fed — siloed information doesn’t compute.
• When AIOps needs a single source of truth for data and control, SD-WAN and SASE architectures provide the foundation — one operating system for centralized information and management.
• In fact, 77% agree that AIOps performs better with a SASE architecture. That explains why respondents rank it as a top AIOps consideration. It enables automation.
With those challenges in mind, you need to ask a couple of crucial questions in order to understand how AIOps could work for your organization. Can your network give AIOps the information feeds and architectural support it needs? If working with a service provider, is their network software defined all across the globe? Readiness hinges on ubiquitous modernization and cloud agility.
Even with the foundation prepared, automation still takes time. Best practices say start slowly, therefore starting early helps. First, use AIOps visibility analytics to fine-tune performance and operational playbooks before attempting to automate processes. Once machine learning has an opportunity to understand your operational parameters, AIOps will be able to show you the right buttons to push to unlock bottlenecks or tighten security. With time, you’ll learn to trust it to fix things for you — without giving it your explicit approval.
While AIOps can’t solve all of the world’s problems, it can create a superior ecosystem for business. Data is mixing and mingling in new ways to make systems smarter, and we’re putting into place the discipline to arrive at full autonomy. When the journey is equally important as the destination, what’s your milestone marker on the way to automation? I’m happy to continue the conversation as the industry trains for this marathon, getting smarter together, one sprint at a time.