IBM CEO Arvind Krishna has a new assignment for Watson, the company’s AI software: working in IT. Best known for its roles as an advisor to oncologists and sometime quiz show contestant, Watson is now being pressed into service diagnosing and solving enterprise IT infrastructure problems, making it one of a growing number of tools in the market for AIOps, an emerging IT practice that applies artificial intelligence to IT operations.
Cut through the hype with our practical guide to machine learning in business and find out the10 signs you’re ready for AI — but might not succeed. | Get the latest insights with our CIO Daily newsletter.
Speaking in an online version of IBM’s annual Think conference on Tuesday, Krishna also announced a broadening of the company’s focus on cloud infrastructure to include computing at the network edge, attracted by the potential market for 5G mobile infrastructure.
“Hybrid cloud and AI are the two dominant forces driving digital transformation,” he said.
IBM’s new initiatives in both areas are built on technologies from Red Hat, the open source software and services vendor IBM purchased for $34 billion last July.
Watson retrains for AIOps
Watson AIOps, IBM’s name for the new iteration of Watson, is built on the latest release of Red Hat OpenShift, a container orchestration platform, so it can run across hybrid cloud environments.
It’s designed to evaluate the swarms of alerts generated by IT monitoring tools when an incident occurs, in order to identify and help fix the root cause of the problem.
“Watson AIOps relies on AI to automate how enterprises detect, diagnose and respond to IT anomalies in real time,” said Krishna.
IBM already has an AI-based IT operations management tool, Netcool Operations Insight, that automatically groups related events and provides context to help solve problems.
Watson AIOps will work with Netcool or other monitoring tools such as PagerDuty, LogDNA or Sysdig, looking for patterns in the structured data of alerts and semi-structured data of log files — but also the unstructured data provided by human beings in incident tickets or through collaboration tools such as Slack and Box, with which it will also integrate.
Those monitoring and collaboration integrations are key, as is the use of Watson AIOps’ natural language processing (NLP) abilities to ingest unstructured data, according to IDC’s global AI research lead Ritu Jyoti.
“A lot of things are happening in the incident tickets, and that is always conversational language, so they are using NLP to extract that and automate some of the level 0 and level 1 classification, addressing and routing,” she said.
IBM is a little late to the AIOps game, she said, but its timing is good is nevertheless: “The COVID-19 pandemic is making enterprises more concerned with resiliency, efficiency and business continuity than ever.”
IBM will have to compete in an increasingly crowded market for AIOps tools alongside the likes of Splunk, Cisco’s AppDynamics, Micro Focus Operations Bridge or Zenoss.
Jyoti also called out HPE as a player in this space, with its InfoSight predictive analytics tool, acquired as part of its $1 billion acquisition of Nimble Storage in 2017.
IBM hopes to make life easier for CIOs by rolling its AI technology into other tools too, including a suite of accelerators for application modernization. The suite comprises Application Containerization Advisor and Candidate Microservices Advisor, which will provide support for developers containerizing or breaking up monolithic legacy apps, and Modernization Workflow Orchestrator, which will use natural language processing to capture knowledge about architectural best practices and advise developers.