It could be argued that no one fully understands what AIOps pertains to now in its aspirational rise within the IT management industry and community. AIOps is a moving target and a term hijacked by Observability vendor marketing. It’s hard to pin down.
But for the most part, we can all probably agree that whatever the form of the solution, the ultimate goal is to augment humans’ intelligence in managing complex software systems of systems (SoS) and (micro)services – highly interconnected with emergent dynamics that can be attributed to rapid rates of change within the environment and the underlying behavioral of components.
If the goal is indeed augmentation and assistance, then a definition of AIOps would start by looking at what precisely human intelligence is. It would be expected that AIOps would share some of the same characteristics of human intelligence.
Note that a tool, technique, or technology could assist service management but still not be considered AIOps. It is not the degree of assistance but the level of intelligence that a machine operates alongside a human operator.
Here are some of the critical characteristics of human intelligence:
– Creativity: imagination, inspiration, innovation, ingenuity, insight, fluency, flexibility, curiosity
– Reasoning: logical, critical thinking, hypothesizing, simulation, prediction, projection, ab|in|de|ductive
– Memory: retention, retrieval, attention, learning, organization, consolidation, planning, forecasting
– Language: contextual, communicative, symbolic, structuring, generative, dynamic, pragmatism
– Perception: awareness, interpreting, distinction, detection, sensing, seeking, recognition, representation
– Adaptability: flexibility, versatility, variety, resilience, resourcefulness, agility, adaptation, feedback
– Problem-Solving: situational, analysis, knowledge, causality, feedback, experimentation, exploration
– Social: mind reading, empathy, attribution, identity, cooperation, collaboration, hierarchies
Looking at the list above, we need to consider each item in terms of the following criteria:
– the degree required for artificial intelligence to operate alongside human operators
– the benefits that can be obtained with augmentation and assistive technologies
– the importance within the context of monitoring and managing software services and systems
Up to this point in the evolution of products and technologies, only Memory and Perception have received attention from vendors, though with little success beyond rudimentary groundwork, because of the interconnections across interrelated cognitive processes.
The question that needs to be asked is whether the standardization efforts in observability can be a foundation for building the cognitive processes for the above intelligence and capabilities we seek.
Many cognitive scientists proposition the importance of language. Yet, no language that supports the appropriate level of communication, coordination, cooperation, and collaboration between humans and machines in managing systems of systems (SoS) and (micro)services has been adopted.
Instead, the engineering community talks of traces, logs, and metrics, which seem incredibly far removed from the abovementioned characteristics.
There is no path of progress from this starting point until there is a course correction.
In a future post, developing a suitable language will be explored to address better the conceptual, contextual, and communicative challenges within the service management domain, such as complexity and change, systems and services, situations and stability, sensing and steering, etc.
ChatGPT is not the answer. Sorry. But its underlying learning mechanism does point to a way forward.