Thoughts on Data Science, IT Operations Analytics, programming and other random topics


Why Analytics are Critical to IT Operations

30 Sep 2015

(first published on IBM Service Management 360 site)

IT Operations Analytics are an emerging area that is increasingly seen as the next big thing in the Operations and Service/Application performance management segment. However, for many of those responsible for performance management, it's no longer a 'next' thing, rather, it is the 'current' thing. If you are not adopting analytics, or at least considering it, now, you are already starting to fall behind. To go a little further, Analytics are set to become so critical to the successful management of your IT systems that if you don't adopt them, then you run the very real risk of losing the ability to control those environments.

Let me refine that a little bit. That these environments we try to manage are getting more complicated is beyond doubt, but what does that really mean to say 'more complicated'? I think of the complexity as being a multi-facetted beast, with aspects related to sheer number of entities involved and the implications for data volume, to the dynamic and rich-nature of the environmental topologies, and on to the business demands for faster access to insights about those environments. All this has to be managed whilst at the same time, reducing costs! Our current generation of performance management tools have quite the challenge ahead of them to stay on top of all of this.

The reality is, they cannot. Major new performance management capabilities are needed to match the growing environmental challenges. Great advances have been made in collecting performance and related data, storing it, and providing access to it through reporting tools. This progress, while a necessary step forward, is not sufficient. It only provides the basis for what we must do next, which is to gain actionable insight about the environments, from that data, Looking at the data the same ways as we've always done, without advances in how we analyze and look at it, using the same old basic reporting will only leave us further overwhelmed. Improvements in collecting and managing the data solve one facet of the complexity problem. To solve the other critical area, we really need to evolve our analysis of that data.

Enter 'Analytics'! What we now call analytics is really just the logical evolution of what we've always been applying to that gathered data ie analyzing it in various ways to gain insight. The difference is that now, that we are carrying out increasingly sophisticated analysis. We've had to move beyond the basic aggregations and other simplistic summarizations, to look deeper into the data, how it evolves over time, how the various sets of data relate to each other and so on. Arguably though, we are at a step-change/increase in our capability where we've gone from a mode of incremental changes in our analytic capability (e.g. a little finer grained aggregation) to dramatically new levels of capability, where we can look more deeply at data within a domain, such as logs, and also looking at data across domains, like combining events and performance metrics. See Combine perspectives to extract more value from your IT data for some futher thoughts on this. These improvements in analytics capabilities have been made possible by advances in the general art of data processing, but also specifically, advances in Machine Learning.

Whilst the technological underpinnings of these analytics may be sophisticated and modern, the objectives of deploying these capabilities are (relatively) timeless. With these analytics, we are trying to solve the same kind problems that we've been dealing for decades, e.g. downtime avoidance, operational cost reduction, customer satisfaction. The difference is that in the face of increasingly complex environments, the intrinsic nature of those problems have become much more difficult. Analytics is just the label we've been applying to that newer class of tooling required to solve those tougher problems and is the key to managing increasingly complex IT environments.

My colleague has written here about gaining competitive advantage with these new analytics capabilities here How IT Operations Analytics Helps you Gain Competitive Advantage, There are many new opportunities created by these Analytics. However, before even worrying about the advantage aspects, you first have to make sure that you are on a firm footing and taking care of those basics. Without modern Analytics capabilities, you run an increasing risk of becoming overwhelmed by the rising complexity of your environments. With suitable Analytics deployed, you at least have a chance of keeping on top of things.

One last thing, these Analytics build upon modern data collection and management regimes. These regimes should already be in place before you attempt to explore modern Analytics. Of course, while you may be in the early stages of Analytics exploration, or even adoption, I'm sure you have a mature monitoring infrastructure in place already...but just in case you don't, IBM can help there too!

I’d love to hear about thoughts on adopting Analytics or indeed Operations Analytics in general. You can contact me here in the comment section below or on Twitter @rmckeown

comments powered by Disqus