As with many words in the English language, the word integrate (or integration) is derived from the Latin word integratus which means to “to bring together or incorporate (parts) into a whole.” The word integration has been in existence since the 1600’s which means that people have been integrating things for a very long time. When I hear that two products are integrated I generally have a positive reaction and assume that the integrated-whole is better than the individual parts. But is that really the case – and is that enough?
After years of hearing the challenges, pains and requirements of IT operations professionals, I have the impression that when people learn that something is “integrated”, they assumed that it’s also (by default) - simple to use and will meet most or all their needs. Now I was inclined to agree with them but then there’s that old adage, “don’t assume anything.”
Many times integration between management products simply consists of a launch-in-context capability, where the operator has the ability to; for example, seamlessly switch from viewing a device in an incident ticket, to viewing the historical performance or capacity utilization of the same device from another application. Now it might be simple enough to swivel from one application to another, but it is worth it? Does this integration (e.g. launch-in-context) give the operator enough information to triage and remediate the incident faster as a result, or does he/she simply have more data? Unfortunately in many cases it’s the latter, which is minimally useful, since it is information that enables decision making and data is simply the input.
So if integration is not enough, then what is enough or required to make IT operations more proactive? I suggest the combination of analytics, visibility and workflow is the key…
Analytics – is a contemporary and very popular word that is at times overused and under defined, but it’s relevance to IT Operations is growing. In the context of IT operations (i.e. managing the availability, performance and capacity of the IT infrastructure) the word analytics equates to intelligence. For example, intelligence that can be applied to vast amounts of complex data, to identify patterns of behavior or correlate business metrics to infrastructure utilization, allows IT operations to identify normal/abnormal activity (and reduce incidents) and understand which and how infrastructure resources will respond to business demand.
Visibility – is something that IT Operations has had for years in the form of availability, performance and capacity charts, graphs and reports. But many of these views display nothing more than data and require extensive analysis to understand where to focus your efforts. However, visibility that’s driven by analytics provides “actionable” views that can quickly identify the hot spots in the infrastructure and help prioritize efforts, speed analysis and remediation. Having multiple levels of visibility is also a must. No longer are device utilization views all that are required. Some views, like a view showing the number of servers in the 95th percentile (2 std. deviations) of CPU capacity used, are required by the majority of performance/capacity analysts. Factor in the professional preference as well as the skill and ability of the analyst and what’s required are views at all levels - infrastructure, applications, services and the data center.
Workflow – is the means in which actions are taken and things get done “automatically” – and are driven by intelligence so the right things get done right (and quickly) instead of doing the wrong things right, or the right things wrong. IT workflows have been around for decades but in my experience there’s certain things that IT operations can and will allow to be done automatically, such as server configuration. Then there are actions, such as automatically adjusting capacity to a mission critical server, that (rightfully paranoid) IT operations folks will only do semi-automatically at most.
We have seen an explosion in IT complexity, speed and scale with the advent virtualization, converged infrastructure and cloud computing. In my opinion IT challenges are only compounding exponentially as these new technologies are not completely displacing physical systems, which come with their own challenges. In short, people can’t keep up and automated workflows, driven by intelligence, are what will reduce human error and increase operational efficiency.
For years IT organizations have struggled to deliver proactive IT operations. While product integration provides efficiency benefits it’s the addition of analytics coupled with visibility and workflows that will propel organizations from simply improving mean time to repair (MTTR), to achieving the higher value mean time between failure (MBTF).
Don’t get me wrong, I enjoy using integrated products – my iPhone being one of them. But sometimes integration is not enough.