Home > The Unifying Force of Data

The Unifying Force of Data

Someday, we may look back on IT as a subset of social science as much as a technological discipline. Because it sits at the intersection of business and technology, visibility and information are at a premium within organizations. In another social science, economics, there is a theory that given perfect information, rational humans will behave predictably. Putting aside the assumption about rationality (and that's a major bone of contention), we can use that same principle to say that when people seem to behave unpredictably, it's a failure of information. In any organization at scale, information disparities have the potential to cause confusion, or worse, conflict. One of the most typical examples of this in the enterprise is between storage administrators and database administrators (DBAs). Very often, these two roles butt heads. But, why? After all, at the end of the day, both are trying to serve the same organization with a clearly defined mission, as discussed in the eBook, The Unifying Force of Data, co-authored by myself and Keith Townsend. So, it could be said that both roles have the same overall goal, but have different priorities within it. These priorities inform how each determines their success within the organization. It's due to a lack of knowledge of their counterparts' priorities that often cause this seemingly inherent conflict. So, what are these priorities? From a storage admin, much of their focus rests on cost. As data continues to eat the data center, the amount, and subsequent cost, of storage, is the fastest-rising cost in a data center. This tends to make them the master of "no" when it comes to storage requests. This focus on costs informs how storage admins interact with other members of an organization. When it comes to DBAs, this can create a vicious cycle of assumptions. If a DBA requests an additional allocation of storage at a given performance tier, the storage admin is naturally skeptical if it's "really" required. The storage admin will look at what the DBA actually used out of their last allocation, perhaps even digging into IOPS requested by an application as part of their calculation. In the back of the storage admin's mind, there may be an assumption that the DBA is actually asking for more than they need at this moment. This might be because of the lag time in provisioning additional storage. Regardless, the storage admin is trained to be skeptical of DBA provisioning requesting. This is where a lack of information can really hamper organizations. The storage admin thinks the DBA will always seek to overprovision either capacity or speed. This causes additional delays as the admin tries to determine the "actual" requirements. Meanwhile, the DBA assumes the storage admin will be difficult to work with, and often overprovisions as a way to hedge against additional negative interactions. This cycle isn't because the storage admin is a pessimistic sadist looking to make everyone's lives miserable. The point of storage in an organization is for it to be used to support the mission. The storage admin must provide applications with the storage they need. But they feel the squeeze on cost to make this as lean as possible. Now it's impossible to expect either storage admins or DBAs to have perfect information on each other's roles and priorities. That would result in duplicated effort, cognitive overload, and a waste of human capital. instead, a monitoring scheme might approach this to effectively allow the two to troubleshoot their issues by effectively correlating behavior up and down the application stack. This kind of monitoring requires the collaboration of both a storage admin and DBA. It may require a bit of planning, and probably won't be perfect from the start. But it might be the only way to solve the underlying information imbalance causing the conflict between the two roles.
Stephen Foskett
Stephen Foskett is an active participant in the world of enterprise information technology, currently focusing on enterprise storage, server virtualization, networking, and cloud computing. He…
Read more