The Defense Department Must Account for its Information Technology Silos
Among many definitions, the Oxford dictionary defines a silo as a process that operates in isolation. In the U.S. Defense Department, everyone works in separate components. Computer silos have proliferated with the availability of a huge number of customized information technology solutions.
The problem with silos is that people in them are unable to work with others on related problems. An expeditionary force is in a silo if it cannot exchange information with other units. The inability of achieve interoperability will impair coordination. Interoperability with units in the same battlespace is critical for any military action. As the Defense Department is evolving into asymmetric warfare, the capacity to break down silos never has been more urgent.
Information technology has the means to break down inter-silo communications. A key is to define silo interoperability in terms of the budget that has been carved out for every unique information technology project.
How do we define a silo? How many silos are in the Defense Department? These questions have bothered me since the end of the Cold War.
At some meetings I attended, general officers used the term “silo” as an explanation for why battlefield interoperability did not work. Silos were found everywhere. Often, one colonel’s silo was another commander’s way of maintaining control over communications.
An enumeration of the Defense Department’s silos showed up in this year’s Office of Management and Budget (OMB) information technology dashboard (www.itdashboard.gov/data_feeds). It identified 2,904 separately funded information technology projects for fiscal year 2012.
Each OMB project had a unique funding source and funding code, specified by the Office of the Secretary of Defense (OSD) comptroller. Each project was tagged by its funding source name and showed total information technology spending budgets for three years. In addition, the OMB identified each project by its funding sources number and by a separate OMB form 53A classification. These tabulations pinned down the dollar definition for each Defense Department silo unambiguously.
That added up to 61 projects, each with budgets over $100 million, with total spending of $16 billion in fiscal year 2012. In addition, 464 projects each had budgets between $10 million and $100 million for a total cost of $14 billion. And 2,379 projects, costing less than $10 million each, were worth a total of $4 billion.
Defense Department program management must keep track of thousands of small projects to ensure they would be interoperable with larger projects and vice versa. The complexity of maintaining synchronization between large projects and small projects imposed an extremely costly burden on overall information technology spending.
If the Defense Department would operate in a totally interoperable software-as-a-service (SaaS) environment that meets the criteria spelled out in various OSD strategic presentations, then theoretically the department could function within only one silo. Individual projects would differ from each other only by different applications and their associated data. However, the one-silo Defense Department will not happen because of too many legacy enclaves. In addition, the silos already in place have multiyear budgets, which makes them ingrown and not easily changed.
The best that the department can do is to collapse silos. It would depend on a handful of shared pools of networking, storage, servers, operating systems, middleware and control commands. The department then could migrate from 2,904 silos to running only a few infrastructure-as-a-service (IaaS), platform-as-a-service (PaaS) and SaaS clouds. Costs would be materially less, whereas security would increase manifold.
However, the Defense Department currently is not set up to proceed with the shrinking of silos. Depending on how each project sets up its infrastructure, the department must continue at this time to operate thousands of silos because that is the way funds are parceled out to program managers. Large projects are burdened with excessive maintenance costs. Small projects keep changing what they need. The results are elongated schedules and excessive costs. Therefore, defense systems are neither interoperable nor effective in meeting emerging military requirements.
The Defense Department must get rid of hundreds of incompatible silos. It is the variety in silo-designed infrastructures that is sinking the department into a swamp from which it must extricate itself. Picking small increments, such as moving a small number of emails into clouds, is useful but falls far short of the need to wipe out thousands of silos.
Paul A. Strassmann is the distinguished professor of information sciences at George Mason University. The views expressed are his own and not necessarily those of SIGNAL Magazine.