6 AIOps trends for 2021

[[386132]]

The increasingly decentralized, heterogeneous and highly dynamic modern IT environment has not only failed to reduce the difficulty of monitoring and management, but has also brought more stringent operational challenges to various organizations with the popularization of cloud computing, containers and microservices technologies. Faced with this realistic problem, IT leaders are turning to AIOps and applying artificial intelligence to IT operations, hoping to solve the dilemma of increasing IT complexity and insufficient traditional monitoring tools.

According to estimates by MarketsandMarkets Research, global IT practitioners are currently purchasing AIOps tools in large quantities. By 2023, the global AIOps market will grow from US$2.55 billion in 2018 to US$11.02 billion, with an average annual compound growth rate of 34%. According to a study released by Digital Enterprise Magazine (DEJ) in May 2020, the number of organizations that have or intend to deploy AIOps has increased by 83% since 2018. "The growing trend of AIOps represents a positive development signal for the future market." Dennis Drogseth, vice president of research at EMA, emphasized.

AIOps does make impressive promises, including delivering fundamental improvements in the speed and accuracy of detecting, diagnosing and resolving IT issues, thereby significantly reducing the chances of disruption to critical applications and digital services. However, the adoption and deployment of AIOps has also become a new challenge in itself.

Below, we look at six trends that IT decision makers should pay close attention to in their AIOps strategic planning for 2021.

Trend 1: AIOps will bring more hype and trouble

Similar to the various technical terms that have been hotly debated in the market, the definitions and understandings of AIOps are also varied and different. According to DEJ research, 64% of respondents believe that the development prospects of AIOps solutions are "confusing." EMA's Drogseth agrees that AIOps is still shrouded in a fog.

Experts say that as AIOps becomes more popular, and each vendor has its own solution design and understanding, IT leaders must find the AIOps products that truly suit their needs from this chaotic carnival.

Gary Brandt, product manager at Micro Focus, said that when choosing the right AIOps tool for your organization, you must invest a lot of effort in careful research and evaluation. "The core challenge for customers is to distinguish which AIOps has real ROI and value and which is just hype," he said.

Lars Rossen, CTO of Micro Focus, said that AIOps will gradually begin to be implemented next year, which also means that some of the hype factors will gradually disappear. "Some people mistakenly believe that AIOps has magical powers that can solve all their problems - this is obviously impossible."

To think correctly about the specific methods of AIOps, we must first understand it as a continuation and expansion of the IT operations analysis technology established in the past 20 years. Brandt believes that whether it is monitoring, event management, problem isolation or automation, AIOps changes the "way" of work. In this regard, AIOps can indeed make enterprises better. IT leaders should also understand their reasons and goals for adopting AIOps. Brandt mentioned, "What improvements do you want to achieve in a hybrid IT environment? Please be clear about the specific goals."

Torrey Jones, chief consultant at Greenlight Group, believes that the most important thing for organizations that are evaluating and implementing AIOps is to clarify how the vendor understands the core ideas of AIOps.

Compared to AIOps, EMA's Drogseth prefers to use the term "advanced IT analytics" because the scope of reference here is no longer limited to IT operations, but also covers DevOps, IT service management, security operations and business stakeholders. In his view, AIOps is not a traditional market, but more of a "prospect", in which vendors represent a variety of implementation methods. Which method is better undoubtedly depends on the actual needs of the organization.

In its recently released AIOps: A Guide to Innovative Investments, EMA outlined a series of core criteria that AIOps products and services should follow, including:

• Ingesting large amounts of data from cross-domain sources

• Access key data types such as events, logs, and configuration data

• Self-learning capabilities to provide predictive, prescriptive and actionable insights

• Supports a wide range of advanced heuristics

• Can cover and integrate multiple monitoring tools

• Support for private and public clouds and hybrid/traditional environments

• Supports multiple use cases

Trend 2: AIOps will add support to security work

AIOps has shown the potential to enhance cybersecurity for years, and 2021 is likely to be the year that potential becomes reality. David Linthicum, chief cloud strategy officer at Deloitte Consulting, said that by bridging the gap between IT operations and security operations, AIOps will further improve system uptime and reliability.

For example, through security data visibility, AIOps can determine that certain application performance issues are not caused by IT failures, but by cyber attacks on the underlying servers. Linthicum emphasized, "In the face of such special situations, we should initiate security processes to implement defenses. However, in terms of traditional tools, such situations are often still regarded as normal performance problems and are rarely associated with security threats."

Linthicum added that this capability would turn operations into the first line of defense, ensuring that managers can shut down attacked servers or compromised storage systems in a timely manner.

Trend 3: Tool suppliers will usher in a wave of mergers

As the AIOps market develops, more vendors will enter the market, and the phase of elimination will follow. At present, many traditional monitoring tool vendors are planning to add AIOps functions to existing products, and the way to achieve this is naturally to directly acquire some smaller AIOps vendors. Linthicum believes that "this will bring better AIOps technology to customers."

As a concept and technology category, AIOps represents the full maturity of operational tools. In recent years, most vendors in the traditional operations and CloudOps fields have begun to "integrate AI engines into their own tools, and whether they can be actually used or not, the results will be called AIOps."

Next up are AIOps tools for startups. These tools are designed with AI technology in mind. Which tool is more suitable also depends on the actual needs of the organization.

Such startup tools tend to be more innovative, use AI technology more effectively, and support modern systems such as cloud platforms. In contrast, traditional tools that have undergone AIOps tend to support legacy systems and mainly serve IT operations personnel who want to manage both traditional and cloud systems with a single tool.

"Of course, the two are complementary, and many companies are using both types to cover all operational infrastructure," said David Linthicum.

Trend 4: DevOps professionals will widely adopt AIOps

AIOps is primarily aimed at IT operations groups, but its appeal is also beginning to expand to DevOps teams, helping them monitor complex environments through advanced tools and generate raw, detailed, and extensive observability data, including logs, metrics, and traces.

By applying AI and machine learning (ML) algorithms to observable and monitoring data, AIOps will become part of the DevOps tool chain, focusing on monitoring and managing issues such as testing, performance and security. Linthicum believes that this approach can also "provide real-time feedback to integrated tools and DevOps engineers through automation."

Nancy Gohring, senior analyst at 451 Research, S&P Global Market Intelligence, said that in the company's recent survey of DevOps practitioners, 42% of respondents defined AIOps as tools that use AI and ML techniques to analyze logs, traces and metrics. "This is consistent with how people define observability. So there is often an interesting intersection between AIOps and observability among DevOps practitioners."

Trend 5: Automation capabilities will be expanded

Since its inception, AIOps products have always considered automation as a core element, but the specific implementation of automation can take many different forms. Although AIOps products have not yet completely eliminated the need for human intervention, this aspect is expected to be significantly improved in 2021.

Today, more and more products are incorporating automation capabilities to solve problems related to use cases or workflows. Micro Focus' Brandt said, "Now that we understand how to trigger or drive this process through analytics, I think automation will be a big trend in AIOps."

According to Rossen of Micro Focus, AIOps is expected to usher in four specific automation upgrades at the technical level and enterprise application level:

• The “typical” category of AIOps automation, which is ingesting large amounts of logs and then analyzing them using ML algorithms to detect anomalies that deviate from baselines.

• Robotic process automation to trigger remediation procedures without human intervention.

• Analyze and correlate topology data to see how connections are made within your system, and then use this information to discover the root cause of problems.

• Automation, which helps end users intuitively resolve issues through intelligent virtual agents (e.g., automated help desks), significantly improving the accuracy and usefulness of customer support.

Deloitte's Linthicum also highlighted the progress of self-healing capabilities. He stressed that some AIOps can already "solve problems discovered by management or monitoring systems." After the AIOps system discovers a problem, it will start the corresponding process to try to solve it, such as restarting a server or network hub. "The main trend is moving towards proactive or self-healing AIOps tools," David Linthicum concluded.

Trend 6: The application of collective intelligence will continue to deepen

Another important trend to watch in AIOps is "collective intelligence," where AIOps vendors aggregate all monitoring data from their customers to create general trends that can be shared with other parties, according to 451 Research's Gohring. "They analyze data from their entire customer base to develop benchmarks and insights," Gohring said.

For example, suppliers can look at performance indicators from customers in specific areas and predict what problems exist based on collective knowledge. In addition, she believes that suppliers can also analyze customers' solutions to specific problems and share them, helping other customers gain valuable insights when dealing with similar problems.

Summary: AIOps will shine in 2021

As 2021 approaches, it will be imperative for IT leaders to use AIOps to address increasingly complex environments that are difficult to monitor.

Yes, AIOps is currently highly hyped and may not live up to the lofty promises made by vendors, but it is proven that the technology can automate and streamline IT operations in the real world.

In addition, judging from the above trends, AIOps technology is constantly developing and improving, and its coverage is beginning to expand to areas such as DevOps and SecOps.

Ultimately, AIOps will be able to help IT teams accomplish their most important and pressing tasks, such as application and digital service reliability, stability, and uptime, which are closely related to business performance.