For those of us in the tech space, you’ve likely heard of AIOps, or artificial intelligence for IT operations, which “involves using AI and ML technologies along with big data, data integration, and automation technologies to help make IT operations smarter and more predictive.” The research firm Gartner recently defined two different high-level categories of AIOps: domain-centric and domain-agnostic.
Domain-centric tools focus on homogenous, first-party data sets and introduce AI capabilities to solve specific use cases, such as network and application diagnostics. Domain-agnostic AIOps platforms combine diverse data sets and data types and synthesize them into insight or action.
Basic Domain-Centric And Domain-Agnostic Use Cases
One typical example of domain-centric AIOps would be an application performance monitoring tool (APM) measuring application latency. The AIOps component would identify anomalies in this measurement, such as when the latency is higher than usual for that time of day. When an anomaly occurs, the APM tool would automatically trigger an alert.
An example of the domain-agnostic approach would be a tool aggregating alerts from several observability tools, such as log management, APM, network performance management and user monitoring. The AIOps component would automatically normalize the alerts into a consistent data model, enrich them with additional context and then correlate them together into logical incidents across all data sources.
Domain-centric tools can effectively resolve a particular pain point that exists today. A domain-agnostic approach is instead more likely to be focused on strategic outcomes for your organization.
The Doctor Is In: AIOps For Diagnoses
There’s a medical diagnosis analogy I like to use because it clarifies the strengths and weaknesses of each approach. Health care vendors are increasingly deploying AI technology in their products. Take a high-end thermometer for example — it could use AI tech to identify anomalies in your temperature, alerting you when it’s too high given your age, sex and time of day.
Similarly, a blood pressure monitor employing AI technology can identify long-term trends in your blood pressure, whether positive or negative. By design, the AI-driven blood pressure monitor handles its own data — aka first-party data — but its domain-centric AI capabilities can be very valuable. If you suffer from hypertension, the trend data can be used to set your medication level or highlight the need for a change in your diet, for example.
Meanwhile, a domain-agnostic approach could rely on correlating multiple symptoms to identify a broader set of conditions. For example, a light fever combined with coughing and skin redness could provide a strong indication of a particular illness. Doctors typically need to review multiple symptoms — aka diverse data sets — in unison to understand what’s going on with your body and arrive at the diagnosis and treatment plan. This is where a domain-agnostic approach can be extremely valuable.
Domain-Centric Vs. Domain-Agnostic: When To Use What
The advantage of using domain-centric tools is that each such tool knows its data set better than any domain-agnostic tool ever could; nobody knows the data better than the tool generating it. Nobody can compete with the granular access a vendor has to its first-party data because the full data set may not even be exposed via an API.
Domain-agnostic tools instead excel at ingesting diverse data sets, connecting the dots and creating the bigger picture. Because many enterprises use several different observability and monitoring tools in parallel, domain-agnostic AIOps tools may be more suitable to derive insights and drive action.
At the end of the day, every IT leader or architect evaluating AIOps tools should ask themselves: What’s my most burning pain? If the main burning pain is related to a very particular use case for a particular data set, nothing is better than a domain-centric solution. But if the problem is how to take high volumes of heterogeneous data and turn it into insight, then domain-agnostic AIOps tools are the best fit.
For many organizations today, AIOps tools are vital. AIOps can provide capabilities across event correlation, root cause analysis (RCA) and incident workflow automation. In the process, they can improve monitoring and service management and help surface incident diagnostics. When used properly, AIOps delivers practical outcomes rather than aspirational goals. In the end, organizations should adopt domain-centric or domain-agnostic AIOps based on their use cases, data diversity and the organization’s own business road map.