Specialization is a great by-product of a healthy economy. Unlike our ancestors, we no longer raise our own sheep for wool, make our own clothes and sometimes we don’t even clean our own house. We have experts to do the things they do best while we devote our time to doing the things we do best e.g. our careers. The same thing goes for an organization.
I recently read a blog about DIY Data Center Infrastructure Management (DCIM) systems and the author pointed out some caveats about homegrown DCIM systems including the level of complexity required to get to the level of commercially available DCIM applications as well as the cost of devoting R&D man-hours to a solution which may prove cost prohibitive in the long run.
I want to add a few other points to the topic of a DYI DCIM system which I observed with datacenters with homegrown DCIMs. First, they are usually ugly. They are like the Africa-shaped tortilla. They may do the trick for a use case or two but it’s not something you want to showcase to your executives, customers and, sometimes, even coworkers. Let’s face it, IT equipment constitutes a huge investment to an organization. Datacenter managers go out of their way to make their facilities efficient, presentable and, most of all, very functional. Why would they manage them with something less than the best of breed? Datacenter managers, proud of their datacenters, often showcase them with well-prepared walkthroughs, pamphlets and videos. Many datacenters have beautiful dashboards on big monitors as you enter the datacenter or in the lobby but I’ve yet to see a DIY DCIM dashboard showcased anywhere. The main reason is that people who put together a homegrown DCIM system usually are good at one or two things but typically lack other skills such as report creation or front-end design.
The second point has to do with future usability and functionality. Building a DYI DCIM is a response action which fills an immediate need. The system does what it needs to and usually nothing more. There is only one client that provides feedback and one point of view that creates the system. A commercially available DCIM system takes in feedback from hundreds of client organizations and builds a flexible system which can be grown into. The DIY DCIM systems usually don’t have user friendly features, expanded functionality, and certainly lack industry supported features like flexible workflows, change management, APIs and OOTB integrations to name a few. Features are usually a result of customer feedback and industry trends. However, DCIM vendors have product managers who are specialized folks whose only job is to identify future needs of clients and engage in forward thinking trying to make customers’ lives better through new product features.
The third issue with DIY DCIM systems is that they are designed and created with the best intentions, but the creators don’t usually remain in the same position, or even company, and the system becomes dated very quickly. In my visits to datacenters, I sometimes find DYI DCIM systems which no one uses anymore because they were created but never updated. They range from systems with very creative uses of excel spreadsheets which served a good purpose, all the way to fancier MS SQL or MySQL databases with web interfaces that remained stagnant because the person who kept it up moved on. Eventually, the well-intended system becomes unused, obsolete and forgotten. Too often these DYI DCIM tools turn into a permanent visitor to the graveyard of homegrown applications.
Creating your own DCIM tool in-house may seem like a good idea. You have the people, systems and resources to create the tool that will provide the features you need. How hard could it possibly be, right? Ponder upon this: even the biggest software developing companies in the world go out and buy a commercial DCIM tool. I wonder why that is…?