There’s a simple reason why most organisations rely on maps and diagrams to plan and manage their IT infrastructure. Humans understand topology and interrelationships visually, and plotting out the various devices, services, and configurations the organisation relies on makes everything easier. But even the most elegant, wall-sized Visio art can hide an unfortunate truth—incorrect maps may be doing IT managers more harm than good.
Here Be Dragons
Traditional IT infrastructure maps face the shortcomings of traditional street directory: going out-of-date. Any manually-derived IT map can only capture as many infrastructural components as the mapper’s skills and time allow. In most instances, IT managers can’t spare much of either: handling the growing complexity of today’s infrastructure is time-consuming enough. As that complexity grows, it further renders manual maps incomplete and unreliable. The solution, of course, is to regularly update the map—but it’s a tougher IT manager than most who can muster up the discipline to do so even once.
As a result, your IT infrastructure maps end up with the technical equivalent of “here be dragons” or “unknown continents”—large pockets of opacity or vague sketches that don’t really reflect the lay of the land. That, in turn, can easily lull IT managers into a false sense of security. Teams with no map at all will at least become vigilant or cautious enough to poke around and discover potential issues. But an infrastructure map with outdated or incomplete information can lead IT managers astray if they’re not careful.
Real-time Cartography
Once out of budget reach for many IT managers, dynamically updating infrastructure maps are now included in many of the management tools admins rely on or are an upgrade away. Maps adapting automatically to reflect shifts in configurations or infrastructural “inventory” also reveal how evolving component relationships influence one another. Automated network-mapping software can even show us potential trouble spots we might encounter if we continue our current trajectory or novel issues that arise out of nowhere and disappear just as quickly.
Most organisations can easily achieve greater accuracy and visibility of their maps by automating rediscovery of new hardware or applications. This can be done within most system and network management platforms, while automated application tracing and analysis of inter-application traffic can help bring interdependencies and less-visible interactions to light. That said, IT managers shouldn’t rely solely on automated routines to brush up the quality of their maps; they should also pay close attention to the qualitative feedback they receive from across the organisation. Anecdotal information and reports can highlight how certain areas of infrastructure may affect one another in ways the most rigorous automated processes might not detect.
Perhaps most importantly, IT managers need to constantly question the natural assumption that they know how their infrastructure works. This isn’t a slur on IT’s competence—far from it. There are so many different types of devices, applications, and platforms today that even those appearing to perform the same function will often do so in wildly different ways with significantly variable results. Even systems vendors claim to be similar will often vary in unexpected ways. When it comes to mapping out those systems and their place in the infrastructure landscape, it’s our job to remember the adage about “assume” and treat every component with sufficient scepticism—exploring the territory until we know what really lies out there.
Charting the Clearest Course
Manually mapping out IT infrastructure won’t get any easier. The move to cloud and customised applications has made it harder to see what’s going on than in traditional on-premises infrastructure. In a similar vein, a growing number of IT elements don’t provide standard discovery interfaces that can easily plug in to the organisation’s map for real-time updates and visibility. Some technologies, like SDN, SDS, and other software-defined controllers, can improve insight into how infrastructures perform and connect to each other. But it’ll still be an upward battle for many IT teams to build self-correcting, true-to-life maps.
Should IT give up on the entire mapping exercise? Of course not. In fact, IT infrastructure maps matter more than ever precisely because of the technical complexity all organisations face today. Today’s networks involve far more stages than before: a single VoIP call, for example, can hop hundreds of times across the network. The latest applications rely far more on outside elements like APIs and the cloud.
The more interlinked our devices, applications, and overall infrastructure become, the more potential points of failure we face, where a single misstep can compromise the customer experience or overall performance. With a reliable map in hand, however, preventing and correcting those missteps becomes far easier and faster. And IT managers already have many of the automation and visibility tools at hand to draw—and keep drawing maps—that’ll let them safely manage any situation.
