As software engineering practices continue to evolve, through this blog we intend to bring forward maturing perspectives and the role of AIOps in this evolution. It is needless to say that in the last decade we have witnessed the evolution of software engineering practices primarily addressing the DevOps cultural revolution. Organizations are taking different paths to adapt to this change, influenced by the number of factors need to steer innovation, faster time to market, tackling workforce silos and other related limitations of legacy practices. As the DevOps cultural revolution results in maturing tools, techniques and practices for DevOps-centric needs, there is a new wave of AIOps-oriented practices in making. AIOps, in other words, artificial intelligence for IT operations.
The size of the prize to move on to AIOps is substantial, for many it can be widening the gap further to catch to the modern landscape of software and can be limiting for future growth. The penetration of AIOps is likely to cause more disruption, let’s explore the real potential of AIOps, how do we start introducing AIOps and scale it for good. Lastly, we would discuss the key impact of AIOps to create a green ICT footprint & be mindful in our approach.
Exploring AIOps impact index for IT
Primarily AIOps, Artificial Intelligence for IT Operations uses data and machine learning to improve and automate IT service management. Another simplistic definition of AIOps is “AI-augmented DevOps practices rely on AI-enabled tools to augment every phase of the pipeline” – Gartner. So, one might ask why do we need AIOps and how to capitalize on the impact index?
According to Arthur De Magalhaes, a senior technical staff member at IBM through a talk presented at DataOps Summit Canada, looking at the last decade’s shift towards digital is compounded year on year. This quick shift has resulted in growing tools, applications and hybrid landscape in the product portfolio, according to some portfolio managers it can be up to thousand applications & tools in one portfolio, and on top of everything major attrition and skill gap. According to the Washington Post, 2021 4.3 Million quit jobs in august, a pandemic-era record.
The factors mentioned by Arthur lay the foundation for AIOps, organization will look at the productivity improvement of the workforce by introducing AIOps in the next five years. Organizations also understand that the complex mix of tools and applications will continue to build up and there is an increase demand for shared products and services. Some of the critical differentiating factors for the portfolio would be availability, reliability, and responding to scaling deliveries to handle dynamic consumption. AIOps when tied to the site reliability engineering objectives can create a differentiating experience, especially when the products and services in the portfolio are always-on, customer insights are critical to detect and diagnose, and error budgets are limited.
Initially, AIOps will likely be mainly focused on enhancing the customer experience of an already existing, maturing portfolio rather than creating new ways. AIOps will steer insights into customers’ consumption patterns, and preferences and improve the overall experience by ensuring the reliability of the products and services in a personalized way. It seems that organizations having a focused approach towards AIOps will tend to capitalize on demand compared to the traditional IT players.
So, are you ready to move ahead? AIOps can be a complex onion to peel if the foundation is not laid it might not pay off. It does just not plug and play, it needs some preparation and due diligence.Going from a reactive to a proactive approach with AIOps is a journey. The first few steps in the process can be as simple as ensuring to standardize the foundational component of the product, making sure to get monitoring and subsequently observability in place. Augmenting Observability with real-time analysis and analytics. Once the initial steps are in place, we could expect AIOps to support a more predictive IT landscape with predictive insights, maturing to more proactive ways of enabling reliability, availability and performance with AI-led detection and prevention of issues.
• AIOps for Enterprise IT
Putting in place AIOps for enterprise IT would be an interesting and challenging objective. Bringing the organizational collaboration, governance and shared vision to leverage AIOps would mean that we have the right data sources in place as input, correlate & logically group events based on topology, prepare to build the metrics that matter to the business, identify and validate anomaly to steer AIOps for decision making. Another important aspect would be democratization, how to leverage AIOps for varied stakeholders, technical and non-technical.
• Which products & Services
Fundamental factors that lead to an outcome leveraging AIOps would be the availability of raw datasets, the ability to collect and correlate data, the capability to apply machine learning models to test and train, and eventually taking action on the proposed outcome.
To seize the opportunity, it is evident that we cautiously choose the early adopter based on the volume, variety, and complexity of IT Operations of a particular set or sub-set of portfolios. It is observed that the early adaptor are cloud-native products with DevOps & SRE teams supporting the onboarding to AIOps. It also seems to be an opportunity for businesses and applications which highly depend on IT, which has gone or undergoing digital transformation. According to IDG, 89% of all companies have already adopted a digital-first business strategy or plan to do so, which creates a bright perspective for AIOps. Lastly, to capitalize on the full potential of AIOps, it might require some time and a step-by-step onboarding of the number of portfolios and correlate them in one way through AIOps adoption at scale.
Full-stack approach for AIOps
AIOps is not a special project, in fact, it adds up to the new equation of the modern cloud-native IT landscape. Then how to visualize IT landscape five years down the line and create a competitive roadmap with AIOps?. It seems to accomplish that goal we would need more than AIOps, it starts at the bottom of the stack with DevOps, breaking down your applications and tools into more cloud-native microservices-based architecture, going to the next steps by embracing DataOps, making it possible for your cloud-native IT landscape to leverage data to continuously improve and innovate with MLOps and eventually maturing towards AIOps at scale.
• AIOps at Scale
Scaling AIOps is a lucrative value proposition for large and complex organizations, Gartner also predicted that “large enterprise exclusive use of AIOps and digital experience monitoring tools to monitor applications and infrastructure will rise from 5% in 2018 to 30% in 2023.” It seems that the fragmented tools in the space for monitoring, observability, and analytics will converge into a more platform-centric approach, making it easier for organizations to correlate and infer data coming from the various systems, at least it has to support interoperability. Another area where an organization invests to scale AIOps is how to collect, curate, and manage a large set of data. AIOps at scale will uncover strategic areas to improve the bottom line over time
• AIOps for Green Technology
In our last blog, we touched upon how DevOps can enable organizations to green ICT products and services, some of the key components would be enabling dynamic provisioning for infrastructure to take a green approach. Another way to look at it is multi-tenancy, producing software with reuse of infrastructure in mind.
AIOps would further strengthen the proposition for optimizing the footprint of IT by AT-led dynamic provisioning, auto-scaling, optimization of IT resources, and balancing the overall consumption. We will explore more ideas through our next blogs as there might be many more possibilities.