In my previous article, I outlined how three major waves in application performance management (APM) have been identified and categorised into the challenges of black box, topology and big data.
These are described as:
Wave One: Waterfall − The Black Box Challenge. This approach from requirement to delivery is said to have been a linear and hierarchical process where it could take years before a finished product was delivered to end-users.
Wave Two: Agile − The Topology Challenge. Agile introduced a much more iterative approach to product development that was less bureaucratic and provided a more predictable timeline for product delivery.
Wave Three: DevOps − The Big Data Challenge. Today, there's just too much data to monitor by hand, which is why the heart of DevOps is automation. It is the key ingredient for building a continuous integration pipeline to improve release speed and quality via Dev and Ops collaboration.
None of this goes anywhere near explaining why APM is part of the AIOps stack and why you need to invest in both – so let’s address that now.
APM Digest describes the origins of APM as predominantly rooted in monitoring. Even where it includes some predictive capabilities, it is said to focus primarily on application performance, including some application/infrastructure interdependencies.
It goes on to add that while the boundaries do become somewhat blurred as APM vendors increasingly embrace machine learning and analytic capabilities in the broader scheme of things, APM tools, given their domain-specific focus, can become critical sources for more far-reaching AIOps platforms.
I would agree with the view that AIOps is a unifying technology that embraces APM, network management, systems management, database management and cloud across multiple use cases, which it does by assimilating and proactively analysing data from a wide variety of sources.
However, AIOps is intrinsically affiliated with advanced levels of automation, making its value proposition more expansive and greater than that of APM. But the point is not to dismiss APM as just a monitoring tool, but rather view it in the broader picture of how it contributes to meaningful AIOps capabilities.
APM as part of the AIOps stack – where does one start and the other end?
APM has an important role to play in the management of applications and the assessment of effective service management and delivery.
Moreover, many APM solutions have developed dynamic capabilities for discovering and modelling application infrastructure interdependencies and discovery and dependency mapping. These in turn have become the foundations for many AIOps solutions, either through direct integrations, or through their own discovery capabilities.
End-user experience management is increasingly relevant to AIOps solutions – it must be in a world seeking to deliver business value.
End-user experience management is increasingly relevant to AIOps solutions – it must be in a world seeking to deliver business value. APM is seen to be putting a growing focus on this area also.
Furthermore, many APM solutions have determinedly sought to embrace business performance metrics which are similarly a growing area of AIOps investment. However, it needs to be emphasised that AIOps platforms are potentially a more versatile foundation for assessing business value, specifically in the areas of capacity, costs, security/compliance considerations and other metrics.
Can AIOps investments unify IT?
The straight answer to this is – yes. AIOps is a unifying technology, and as such, a platform for change potentially across the full IT spectrum in any enterprise and not just where its name implies; ie, operations.
Application architectures have grown significantly in vertical and horizontal complexity. Next-generation monitoring solutions must now deal with dynamic applications which depend on mobile/IOT devices, hybrid cloud deployments and legacy on-premises deployments. Additionally, modern applications are based on small, distributed containerised microservices doing small units of work.
Therefore, IT is a diverse, complex and constantly changing landscape where manual processes to diagnose problems before they arise and put prevention measures in place via manual processes is simply no longer viable.
The growing complexity of the current application landscape has created a ‘big data’ challenge where manual processes can no longer scale to arrive at root cause in a timely manner.
A bad user experience on a revenue-generating application can damage a brand in minutes and result in loss of business – customers today have many choices which they exercise digitally and swiftly. Loyalty needs to be earned through ‘consistent’ exceptional service – the second that ball is dropped, the customer moves on.
The current and third wave of application monitoring must address the ‘big data’ challenge through the utilisation of machine learning algorithms to perform anomaly detection, pattern recognition and predictions so that issues can be highlighted and remediated automatically – essentially pre-empting production problems and fixing them.
Getting IT to align with business goals
This is where AIOps comes into its own. It enables the support of online operations, business application owners, digital transformation teams and enterprise operations, as well as non-IT business executives.
It achieves all of this in part through business-driven service analytics platforms that provide business service-based views across the full application infrastructure.
These views are informed by critical key performance indicators, such as: revenue, business activity management, business process efficiencies, conversions from competitive Web sites and customer behaviour.
It also aids in measuring capital and operational expenditure efficiencies across the IT landscape.
If you think that’s impressive, tune in for the final article in this series, which will strive to reveal the AIOps value proposition for enterprise and public sector markets and how it is the future of APM.
Share