Big data, and big data processing is changing and inspiring change in all areas of IT Operations, including Network Operations (NetOps). There are many important benefits already being provided, and on the drawing board to be provided in the future, through AIOps (artificial intelligence for IT operations). However, the more I look into AIOps, the more I appreciate the central role of people. Which is why we are currently living through a period of augmented operations.
Previously I have written about a three-olive network architecture framework. Expanding on the operations part…
Figure 1. Simplified view of AIOps pipeline in operations / NetOps
…If we take a (over) simplified view of an AIOps pipeline, it inevitably comes to a point where a known correction can be triggered to an automation system, or not. If not, then an alert has to be raised to network operations workers. Whether the network operations worker will actually notice the alert amongst all other activity, is a subject for another time. Let’s assume the alert is actually observed and network operations has the intent to take action.
To say that the human brain is a neural network with deep learning capabilities is an understatement, at best. The human brain retains pleasure, trauma, history, experience and generates intuition; intuition that is especially useful in the context where there is great familiarity with the problem domain. Network operations people understand that two devices, either from the same vendor or from different vendors, may behave differently, in the presence of certain conditions, and even in response to similar configuration commands / directives. Humans understand the financial, political, and relationship consequences of various conditions and actions; even if imperfectly.
There are two questions facing operations teams. One, what can automation systems actually do today, and what can automation systems be trusted to do today. Two different questions. The answers to both of these questions will evolve over time, as AIOps / automation improves in its understanding of complex systems, and as they demonstrate competence in dealing with a wider range of issues.
During the epoch of augmented operations, where human agency remains critical, an important challenge faces operations people. The number of operational tools and the distributed nature of human wisdom. This is why top analyst firms from Gartner to Forrester are highlighting alert consolidation/noise reduction and importantly, collaboration, as key aspects of AIOps.
Collaboration harnesses the experience, intuition, and carbon-based neural networks of human teams, in the diagnosis and resolution of complex issues. The best AIOps solutions are literally, another member of the team.
Today, we are at the early stages of AIOps, and while it is already delivering great value, removing noise, amplifying signal, and triggering some automation use cases, people remain a critical part of today’s operations environments. Leveraging human intelligence in operations environments, MUST be an important focal point of AIOps.