Why AIOps Implementations Require an Integrated Approach

AIOps platforms utilize big data analytics, machine learning and a variety of other algorithms to embed IT systems with the ability to analyze data sets, identify patterns and make autonomous decisions. These solutions have seen steady uptake over the past two years, and Gartner analysts estimate that roughly 50 percent of enterprise organizations will be actively using AIOps by 2020.

As with any emerging technology, AIOps currently suffers from some confusion in the marketplace. One particular issue is the notion that you can gradually acquire multiple single-purpose AIOps tools and progressively integrate them into a functional platform. That strategy will likely produce disappointing results.

The goal of an AIOps implementation is to collect and integrate operational information into a common platform, and then use analytics and machine learning to identify, respond to and report on IT issues in real time. However, problems can arise if multiple tools are collecting their own data and storing it in their own databases. If one tool is analyzing data that has already been filtered by a different tool, the resulting metrics won’t be entirely reliable, making it difficult to pinpoint issues quickly.

Gartner analysts say it is important to understand that AIOps refers to multilayered technology platforms rather than a collection of individual tools. The firm has identified 11 distinct capabilities that an integrated AIOps platform should support:

• Historic data management that maintains historical information from across the environment
• Streaming data management that maintains real-time (as opposed to historical) data
• Log data ingestion, wire data ingestion and metric data ingestion that capture information from various sources and prepare it a for analysis
• Document text ingestion that captures, indexes and analyzes human-readable information
• Automated pattern discovery and prediction that analyzes real-time and historical data to detect patterns and correlations that can be used to predict future events
• Anomaly detection that compares system behavior to baseline data to identify departures from normal behavior
• Root-cause determination that leverages pattern discovery and prediction to automate the identification and resolution of common IT issues
• On-premises or Software-as-a-Service delivery that offers the flexibility to deploy the AIOps tools as an in-house or cloud-based solution

To date, few vendors offer comprehensive, integrated AIOps platforms. Of the 20 vendors assessed in Gartner’s August 2017 AIOps market guide, only four covered all 11 capabilities. They were HPE, IBM, ITRS and Moogsoft.

Where Moogsoft stands out is in its ability to integrate and work well with many of the legacy monitoring tools organizations already have in place. Other vendors tend only to support tools from their own platform/ecosystem. Moogsoft was designed from the ground up to support solutions from a number of vendors, including Splunk, Dynatrace, AppDynamics, SolarWinds and more. By connecting with myriad tools and correlating alerts across all these data sources, Moogsoft provides a holistic, single source of the truth.

Given their ability to collect massive amounts of data and analyze it rapidly, comprehensive AIOps platforms can produce a host of operational and economic benefits. A platform-based approach that integrates multiple data sources and tools can help organizations maximize the value of their AIOps investments.