The Moogsoft AIOps Symposium, in San Francisco and — for the first time — in London shows that Moogsoft is continuing to push the boundaries of a rapidly evolving market.
When Moogsoft put on the first AIOps Symposium in 2017, the question that we heard repeatedly was “AIOps? What’s that?” Now in 2018, the term has become part of the lexicon. This mainstreaming of AIOps was shown in the attendance figures this year, with hundreds attending sold-out events in San Francisco and London.
Still, I would argue that while the term is heard more frequently than it was a year ago, there is a lag in understanding of what AIOps means. As with emerging technologies, AIOps and “Observability” tend to be misunderstood as just an extension of the existing category of monitoring.
Moogsoft CEO Phil Tee was at pains to call out this attitude in his introductory keynote. The many and various examples of stupendous quantities of data being gathered by enterprises made a poor contrast with the very few useful analyses which were being done as a result. One of many memorable analogies was that of data as a Ponzi scheme, based on the illusory hope that if only enough data could be gathered, surely enlightenment in the form of business value and insights would result. Instead, in the absence of new techniques to take advantage of the data avalanche, the reality is that more data will lead to less, not more, understanding.
More Data = Less Understanding
Other speakers backed this view, notably Rüdiger Schmid of Daimler AG’s connected vehicles unit, and Natalia Jojić-Ferguson of WorldPay. Both are current Moogsoft users, and were able to speak about the scale of the data problem in their very different industries, as well as illustrating how Moogsoft is helping them drown out the noise and prioritize what’s really important. Their presentations showed how IT has moved out of the back-office to become an integral part of the fabric of our lives, invisibly embedded in the cars we drive and into every purchase that we make.
Alerts are expensive — or rather, wasting expensive specialists’ time on alerts is expensive, both in direct costs, and in the indirect impact of alerts that are missed in the middle of an event storm, and which go on to cause outages to end users.
Vivek Bhalla of Gartner further reinforced the centrality of AI to the future of IT Operations by showing some of the negative consequences of failure to adopt these new techniques. In particular, Gartner believes that 30% of IT enterprises will become non-viable by 2022 due to failure to implement AI as a core part of their business.
From Monitoring To Observability
This need for new algorithmic techniques to make sense of performance and availability data has been a part of the AIOps message, but the AIOps Symposium is also an opportunity for Moogsoft to show how we are continuing to advance the state of the art.
Much of the recent conversation in the monitoring field has been around the notion of Observability. The formal definition of observability involves deducing the state of a system from outside observation of its visible behavior. In other words, we are moving from the old methods based on polling to new techniques built around the assumption that streaming data is abundant and that the question is what is actually interesting and relevant.
After all, alerts are expensive — or rather, it’s wasting IT’s time in terms of direct costs, as well as in the indirect impact of the alerts that are missed in the middle of an event storm, and which go on to cause outages to end users. The problem is further complicated by the move to distributed systems, which are inherently always degraded to some extent. This degradation is not a failure, especially as it is usually addressed through automation. In fact, it is a sign of the system working as designed.
This means that a fundamental shift in the conception of monitoring is required, from the old incident-driven model (“Something is happening! React!”) to a more insight-driven approach, in which the system will intelligently inform the right people of information that is relevant to them, and do so in context.
Introducing Moogsoft Observe
The latest move to make this vision a reality was the announcement of Moogsoft Observe. This newest addition to Moogsoft’s existing portfolio moves centralized analytics outwards to the data source, extending the core AIOps platform capabilities to provide IT teams with absolute visibility and total observability into any customer-impacting problem. Phil Tee demonstrated the Observe capabilities on stage, showing how it can ingest time-series and metrics data in real-time and apply AI to detect problems at the source.
This type of approach is the only way to control the cost of operating modern IT environments. Already today, OpEx is six times CapEx; in other words, companies are spending a lot more time managing IT than actually using IT to deliver on their business objectives. Paul Ferguson of Amazon Web Services backed this up in his presentation, discussing the ever-increasing complexity of cloud-centric application infrastructures, before illustrating how Moogsoft AIOps and Observe integrate tightly with Amazon services to bring visibility and control.
“Duty Now For The Future”
It fell to Will Cappelli, CTO EMEA and Global VP of Product Strategy, to close out the formal presentations, talking about what needs to happen next. He began by sharing his theory about why this fundamental transition has not happened yet, which he identified in Thucydides’ Greek Triad: fear, honor and interest. After a whirlwind tour of his expectations for the next few years, he closed with his “Duty Now For The Future,” in a nod to Devo.
What better call to action could there be than that?
About the author
Dominic Wellington is the Director of Strategic Architecture at Moogsoft. He has been involved in IT operations for a number of years, working in fields as diverse as SecOps, cloud computing, and data center automation.