It’s clear that investors believe in the market potential for AIOps if Dynatrace’s recent IPO is anything to go by. Its shares rose 49% in initial trading, valuing the company at over $6.5 billion. Dynatrace started life as an Application Performance Management vendor. Now, by incorporating AI capabilities for automation and analysis it has broadened its scope into infrastructure management and user experience management.
I want to focus this post on infrastructure management. Dynatrace is by no means the only company who claims to offer AI-fuelled, hybrid-cloud infrastructure monitoring and management. Indeed I recently counted over 60 vendors who claim to do the same thing. The reality is that they have all come from different backgrounds with a different functional focus. Some have been focused on helping DevOps practitioners or have been out and out Application Performance Management vendors. Others have come from a network management background and some have a heritage in on-premises legacy server and storage management. The likelihood therefore is that, while there may be some overlap, end users will still need to work out which vendor is most suitable for their particular requirements.
Given that functional product descriptions and marketing positioning can look very similar, choosing the right tools to evaluate can be quite difficult. We know from research that most enterprises, of any size, have multiple point solutions managing their IT infrastructure. 18 months ago, my colleague David Norfolk and I wrote a Bloor Spotlight, “Instrumenting the Virtualised Data Centre for Performance based SLAs”. We identified the key requirements for providing real-time monitoring and performance management of IT Infrastructure. At the time we felt that there were some emerging solutions that got over the problems caused by siloed, point solutions, in an on-premises environment. However it was clear that gaining visibility and control across a hybrid infrastructure that included public, multi-cloud and hybrid cloud environments was not yet a practical reality.
Dynatrace and others now offer the ability to monitor and manage the whole of the infrastructure stack in a public and hybrid cloud environment. So, progress has clearly been made, but it is unlikely that all customers, or indeed the majority of them, will be able to ditch all their point tools and realise the claim of having a “single pane of glass” to manage an entire infrastructure. If nothing else, the ability to manage both legacy on-premises and public, multi and hybrid clouds through a single pane of glass doesn’t appear to be widely available.
Another challenge comes for customers for whom being able to get business-oriented performance, rather than availability-based SLAs, is a critical requirement. Think financial trading systems, on-line retail applications, ad-serving systems and the like. This requires monitoring systems that collect real-time data at a more granular level than that provided by event logs, for example. Such systems need the ability to identify, and possibly remediate, emerging issues autonomously, before they impact the business application. It isn’t always clear which, if any, solutions can provide these functions.
To be fair, not all customers need that level of granular, real-time monitoring and management. So, Bloor will be undertaking some primary research to evaluate the functionality of the vendors of infrastructure monitoring and management tools. We intend to plot the results against different business requirements for handling scale, business transactional complexity and the impact of poor performance or loss of service. This is designed to help vendors position and articulate their offerings more succinctly, and enable end-users more accurately decide on a list of suitable vendors to evaluate. We’ll be sure to let you know when we publish the results.