What is the future of AIOps? Whether you benefit from IT Operations teams and tools or you’re lucky enough to manage it all (a hint of sarcasm), the IT Ops space is extremely difficult to keep up with. There’s the collaboration layer with ticketing, notifications, and chat. There’s the data layer that includes change management tools, monitoring tools, and topology services. Each of those sections within the 2 layers have a whole host options, not to mention types of tools, specifically in the monitoring section. For example, within monitoring, an IT team typically has between 5 and 10 tools (i.e. APM, NPM, SIEM, ITIM, DEM, etc.) that overlap in many areas but none of those tools are the one tool to unite them all. All of these tools are necessary to manage an ever evolving cloud, infrastructure, and applications landscape. The problem is: how do you manage all of the noise generated from all of these amazing tools? People? More Tools? Less tools? Whiskey?
How Did We Get Here?
In the good ol’ days of IT, where all of the infrastructure and applications literally lived within four walls of a company, IT Operations was simple. Ticketing tools, primarily made up of an ITSM (IT Service Management) platform, such as JIRA or Remedy, would be monitored by incident managers who would get notified through email or a call or showing up at their desk by Betty in finance who couldn’t login to the ERP (Enterprise Resource Planning). The incident managers would investigate and remediate. Done and done.
Fast forward to today. Technology accessibility migrating to the cloud continues to evolve. The availability and commoditization of bandwidth and connectivity is increasing heavily with fiber, wireless (LTE & 5G), satellite, shinier tin-cans and string. The business is asking for more flexibility, increased speed-to-market, and greater geographical dispersement. With all of the technology control being pushed beyond the company four-walls, more tools and more solutions are needed to automate, deploy, manage, support, change and monitor the operations.
The Agent Smith Approach
Historically, the answer was, “Throw more bodies at it!” using what I like to refer to as the Agent Smith approach (no matter how much Agent Smith multiplied in Matrix Reloaded, he could never add enough bodies to resolve his problem, Neo). Since we can’t throw enough bodies at the problem, the market expanded the technology sprawl by adding more supporting tools and tweaking old tools to do more. More expansive ITSM platforms, such as ServiceNow or Cherwell. Supporting tools for incident managers, such as notifications (i.e. PagerDuty) and chat (i.e. Slack or Teams). Change management tools (intended for CI/CD), such as Jenkins or GitLab. More monitoring tools (APMs, NPMs, etc.), such as Splunk, AppD, Datadog, ThousandEyes, and a whole host of others. Typically, enterprises will have over a dozen tools along with all the noise and alerts that come with those tools. Because of the noise, an IT Ops meantime-to-resolution (MTTR) is truly difficult to manage and determine.
Is AIOps the Easy Button?
So what is the answer? It wasn’t more people or more technology (kind of). IT Ops needed a new layer in the architecture – Event Correlation & Automation. Queue the music… enter AIOps, which is made up of Big Data and Machine Learning. AIOps cross-correlates the alerts and data from ALL of your existing monitoring, change, and topology solutions to reduce noise, find root cause faster, and help automate action and remediation. Fortunately and unfortunately, there are multiple suppliers that interpret AIOps differently. Some are decoupled and agnostic from the collaboration and data layers, such as BigPanda. Some treat AIOps as an add-on or a module within core technologies found in the collaboration and data layers, such as ServiceNow or Splunk. As we vetted the future of AIOps landscape, we outlined some significant pros and cons of what is out there today.
3 Tree Tech will be delivering what we found through a series of articles known as “Future of AIOps… What’s Real & What’s Bull?” Stay tuned as we dispel the confusing lanscape of AIOps.
Eric Skeens is the co-founder of 3 Tree Tech in Portland. He is a platform-agnostic tech researcher that transitions siloed organizations into automated DevOps centric businesses. Message him right here.
IT Stuff To Your Face!
For more smarty-pants research, insights, and invites to our CIO/CTO virtual wine tastings, lunches, and scotch sessions, sign up!