Our Reports


Are We There Yet? Autonomous Networking and the Rise of AIOps

Author: John Burke, CIO and Principal Research Analyst

Digital transformation demands a continuously available, high-performing network and drives ever more business-critical activity to the network. Yet most organizations aren’t increasing the size of their network teams. So, IT must increase use of automation as a “force multiplier” for staff.

IT needs automation to help understand what is going on in the network, and also automation to act in response to changes in the network. IT teams are often leery of tools that take action automatically in response to events in the environment, often preferring systems that keep them in the loop. Being given a button to push and the choice of whether to push it keeps control squarely in the hands of staff, and prevents machine-speed mistakes—resulting from a lack of contextual understanding—that expand damage rather than healing it.

AIOps is the application of AI and analytics techniques to the problems of network management. AIOps tools have the potential to act on behalf of IT staff more flexibly, reliably, and correctly than previous generations of tools because they can apply contextual understanding to decisions about what is happening and what to do about it.

Artificial learning and reasoning coupled with self-modification provide the equivalent of digital twins of the network administrators and engineers able to watch, warn about, act on, and improve the network as a driver of business value.

IT professionals should:

  • Assess their, and their networks’, readiness. Can the network give an AIOps tool the information feeds and architectural support it needs to be effective?
  • Assess the kind of AIOps tool—standalone, or built into a management portal for key infrastructure or network services—that will work best for their network given the ways in which it can support integration (e.g. via SDN and SD-WAN controllers, or device APIs, or CLIs, etc.), and where their greatest staffing and performance challenges lie
    • If working with a service provider, how well is that provider’s infrastructure tailored to provide the AIOps tool all the data it needs, and the ability to act effectively on the enterprise’s behalf? How long has the tool been in production use, and how many years of operational data have gone into its training?
  • Start slowly, and early: use AIOps for visibility first, and expect to take time to teach it their context before looking to it for automated responses (hence the importance of starting early)
  • Use the analysis AIOps provides to fine tune their own operations playbooks before attempting to automate them
  • Learn to trust: Lay out a timeline for moving from “show me the button to push” to “tell me you just fixed something” with a rising level of importance over time (and plans how to fall back a level if a step pushes past the tool’s abilities).

Table of Contents
  • Executive Summary
  • Why IT Needs to Stop Managing the Network
  • AI: Automating the Understanding
  • Closing the Loop: Moving from Understanding to Acting
  • AI0ps: A New Generation of Automation
  • Toward the Self-Driving Network
    • Help Wanted: Virtual Assistant
  • What AIOps Needs
    • Data, Data, Data
    • Power to Process
    • Pulling Strings
  • What IT Should Look for AIOps to Deliver
    • Understanding
    • Proactive Focus and Root Cause Analysis
    • Process Support and Integration
  • What AIOps Can Do for the Business of IT
  • Conclusions and Recommendations


You are currently viewing a preview of this content. Nemertes Clients, please log in for full access to all research content. If you are not a client, please click below to purchase access to this research report. We also invite you to become a client.

 



Purchase Access

$1,995.00