The complexity of some IT systems has made it near impossible to understand exactly what is going on. AIOps helps cut through that noise.
The complexity of some IT systems has made it near impossible for workers to understand exactly what is going on, if anything. To cut through the noise, a new category of IT operations, called AIOps, has seen major traction in the last few years.
In a webinar, Will Cappelli, CTO EMEA and Global VP of Product Strategy at Moogsoft, laid out the five questions that AIOps aim to answer:
1. Is something happening?A modern IT system invariably creates lots of data, most of which is redundant or noise. “It’s only when you’ve got to that basic solid core of data that’s telling you what is taking place within that digital environment that you can begin the appropriate analysis and draw conclusions from the observations you’re making,” said Cappelli. To identify high-information content, Moogsoft uses a set of ‘entropy’ algorithms, to remove all the low-quality content from the stream.
2. What is that is happening? After the initial sort, the volume of data may have decreased by as much as 95 percent, according to Cappelli. Data can now be correlated into different sets, using a machine learning algorithm to discover patterns. This can be based on time, topology or data contained within.
SEE ALSO: How APM and AIOps Help Enterprises Cross the Big Data Skills Gap
3. Why is it happening?Correlation does not equal causation. The next step is to identify why something is happening on the system, using inference algorithms.
4. Who should care? After identifying the issue and understanding the reasons, the AI should then figure out who needs to know about this. Having an algorithm that supports data virtualization, natural language processing, and collaborative workflow makes it much easier to inform all those in the organization who need to know.
5. What should be done about it? The AIOps can then provide suggestions on how to fix the issue. From this, IT experts or business leaders are in a much more advanced position on how to react to issues and have more information to solve them.
“When Moogsoft thinks of AI, we don’t think of it as one thing or as being a vague set of properties, we think of it as being able to answer those five questions, by five distinct types of algorithms working in a choreographed way,” said Cappelli.
By supplementing AI in many traditional ITOps stages, it can assess the data at a far quicker pace and leave the workers to come up with the strategy to fix problems. This could reduce the amount of time a system is down, by spotting the issue quicker and identifying the cause.
