Micro Focus is now part of OpenText. Learn more >

You are here

You are here

The state of AIOps: What IT can expect in 2020

Esther Shein Freelance writer

AIOps is becoming the go-to technology for unifying monitoring capabilities because it presents a single source of the truth, according to industry observers. In fact, some 68% of the almost 100 IT leaders that participated in the inaugural meeting of the AIOps Exchange forum said they have active AIOps projects underway.

Enterprises continue to struggle with monitoring their portfolios, the exchange's survey showed. The three top monitoring problems included too many disparate, siloed tools; lack of coverage of their entire IT environment; and an over-reliance on rules-based solutions.

Half of the respondents that already use AIOps said they do so to improve customer satisfaction and experience.

AIOps tools and platforms apply machine learning to IT operational data to help correlate alerts from multiple tools, aid incident management teams in identifying root cause, and use past histories to recommend courses of action. This is according to Isaac Sacolick, president of the digital transformation consultancy StarCIO and the author of Driving Digital.

Automation in IT is not new, but today's AI-based automation has been enhanced. IT departments employing off-the-shelf AI tools are seeing a reduction in unexpected downtime from mission-critical systems. Using AI tools lets IT operations resolve problems within minutes instead of hours and is transforming the customer experience both for IT and the business.

Here's what to expect for AIOps in 2020 and beyond.

In with the new

AIOps remains a relatively new concept for CIOs and IT leaders—​but they are definitely familiar with one of IT's major pain points: how to make sense of the enormous amounts of operational data coming from different tools and systems to diagnose and resolve IT issues most efficiently, said Sacolick. "This is particularly challenging for CIOs managing large-scale hybrid cloud and hosting different applications that require high performance and availability," he said.

AIOps will evolve from a pure-play category to more of a tech-enabling category, with platforms that look different and have different purposes, said Dennis Drogseth, vice president of consultancy Enterprise Management Associates (EMA). Some vendors are focused on the application performance management space, while others will concentrate on security, cloud, DevOps, and ITMS. 

It's still early days to know for sure how the platforms will evolve, Drogseth said. But it's clear that AIOps is an enabler for helping with "what's turning out to be a variety of platforms with different levels of complexity."

Putting AIOps into action

Dealing with system complexity was the goal at Zebra Technologies, which manufactures printers and ruggedized mobile devices. Thomas Eurick, supervisor of service delivery operations, manages operations teams that oversee and monitor all of Zebra's tech platforms throughout their lifecycle. A lot of the platforms are cloud- and container-based and new to the company, but they also run on on-premises systems as well, he said.  

Some of Zebra's newer software systems include RFID, point of sale, and IoT and are designed for retail chains and transportation companies. Eurick began looking at new, SaaS-based monitoring tools because he has a small team. They wanted to focus on using the solution—​"and not have to take on the additional burden of maintaining the solution's infrastructure and updating it," he said.

Because IT is using a mix of new and older network devices and other technology, Eurick said they needed capabilities not found in their existing tool, such as SNMP trap ingestion and support for Docker, Kubernetes, and all three major public clouds.

The idea is to use AIOps as an emerging feature set to better manage events such as network outages or delays. 

Zebra  deployed an AIOps platform on the public cloud and let it correlate some events. Right now, "we are in applied mode for certain groups of resources and observe mode for others," Eurick said. IT hasn’t done anything "aggressive" with AIOps yet, but he plans to more deeply tune and leverage the system heading into 2020.

Eurick says the inference model he's using will pay dividends down the road. Zebra's AIOps platform "has squashed some alerts floods, so rather than getting 50 alerts, it has correlated them into one. When I have something broad going on in multiple services, it reduces the noise."

Where AIOps shines

AIOps is ideal for use cases where IT wants to look for anomalies within a lot of data, EMA's Drogseth said. Vendors will have many different approaches for doing this; some will focus on microservices and containers, some on cloud, and some platforms will support mainframes where others won't, he said.

"The use cases I'm seeing for AIOps are not limited to operations and certainly include IT service management teams and potentially even business stakeholders with business outcomes."
—Dennis Drogseth

The use cases identified by the AIOps Exchange survey include breaking down organizational silos; modernizing the toolset to address proactive, predictive, and reactive scenarios; reducing mean time to detect (MTTD) problems; remediating; and doing root-cause and problem management.

Most organizations want the same things, Drogseth explained: a tool that is highly deployable, provides quick ROI, does not require a lot of overhead, and can integrate with multiple monitoring tools.

He also refers to AIOps as advanced IT analytics, which incorporates cross domains, machine learning, broad data collection, and serving multiple stakeholders.

Going forward, automation will be key, Drogseth said. In 2020, users can expect to see more platforms that integrate automation with AIOps as well as with their existing investments in ITSM platforms and other IT processes.

An effective combo: AIOps with DevOps

DevOps is also a growing area of focus for AIOps, said Drogseth; StarCIO's Sacolick agreed that DevOps teams want clear visibility into the health and performance of applications.

Among AIOps Exchange participants, the influence of DevOps on the purchase process is rising, with 28% of survey respondents indicating that DevOps drove their decision to deploy AIOps. However, the report noted that a blend of AIOps with DevOps requires a culture shift for the latter "because coders have never experienced the degree of exposure to flaws in their apps so early in the process."

By working together, the exchange report suggested, "AIOps can provide DevOps with tools and visibility to maintain responsibility for their components."

Next year, Sacolick said, more CIOs will pilot and evaluate AIOps platforms.

"CIOs are naturally skeptical of machine learning. The early adopters will plug AIOps platforms in and evaluate where they can make an impact."
—Isaac Sacolick

Keep learning

Read more articles about: Enterprise ITIT Ops