The building blocks of true decision support
Growing data streams need constant attention
In all areas of defense, data is streaming in as fast as its systems can accommodate it. Much of this data is unstructured — e-mails, PDF files, recorded conversations, data from sensors, and geographical and positioning data, among other types. This type of data, often called big data, is growing at an unprecedented pace — much faster than traditional, structured data such as data in relational databases. The explosive growth of this data, combined with the need to make accurate, real-time decisions, has resulted in something of a quandary for the defense community. Without these capabilities, important data can be missed, resulting in erroneous decisions.
“The issue is whether it’s possible to create business analytics tools that can analyze massive amounts of data and give you real-time decisions in a world where time is approaching zero in terms of how much time you have to make a decision,” said Jim Flyzik, currently president of The Flyzik Group, former senior advisor in the White House Office of Homeland Security and vice chair of the Federal CIO Council.
It’s possible today, but probably not with one far-reaching system. Instead, it involves a set of technology tools and processes that work together to produce the desired result.
The first building block is some type of business intelligence system that can identify, aggregate and analyze data from all relevant sources. Their main benefit is that they consolidate multiple data sources and formats in one place, allowing for better and faster analysis. These systems generally have customizable rules engines and dashboards, sophisticated query and analysis tools, and a full slate of visualization capabilities. Also very useful are systems that include more collaborative components, big-data analytics for unstructured data and predictive analytics.
The use of business intelligence for decision support is invaluable. Imagine, for example, a mission to identify potential online threats to the United States. Doing that effectively means identifying and analyzing large amounts of both structured and unstructured data, gathered from a variety of sources that might include video, text messages, e-mails, phone calls, credit card purchases and airline records. The only way to analyze this type of disparate data is with a business intelligence system, which can use sophisticated algorithms to link data and identify candidates.
Another important tool extends visualization capabilities far beyond what traditional business intelligence systems provide. This is critical in situations when decisions must be made quickly and accurately. The system should have full 3D visualization to build working simulations, and work as easily on mobile devices as on desktops. With these capabilities, analysts can model the workflow and behavior of potential scenarios.
Along with visualization on the front end, decision support systems should have geospatial mapping capabilities on the back end. These systems, which accommodate base maps, imagery data, demographic and other data, allow analysts to quickly create, share and analyze interactive maps.
Continuing the example from above, consider how useful it would be to use visualization technology to overlay ATM recordings and other data gathered by the business intelligence system with technology that could develop 3D simulations and interactive maps that pinpoint the trail of potential criminals. With this information, analysts could quickly determine which of the scenarios they have developed could be considered credible threats.
Without a project management function, decision support would be nearly impossible. A lack of project management tools can easily result in miscalculations, activity delays and time loss. On the other hand, a good project management tool allows everyone involved in the mission to review activities in progress and perform what-if scenarios to find the best alternatives, and leaders can access project status quickly. With the right tool, project forecasting will be more accurate, tasks will be more automated and time management will improve. The project management function should be integrated with other decision support tools.
Think about how difficult it would be to coordinate all of the pieces of the mission to identify and catch online threats without project management technology. Communication and timeliness are essential to the success of this type of mission, and project management tools are the best way to ensure that everyone involved is up to date at all times.
Other important features and tools in a comprehensive decision support framework include full mobility, enterprise search, configurable workflow, and connectivity with relevant external systems, such as data warehouses and data marts, and financial, planning and forecasting systems.
All of this is best tied up in a Service Oriented Architecture (SOA) — basically, a set of interoperable services built as software components. Tying these technologies together with SOA creates a set of predefined objects and components that are already built. It also eliminates the problems and time lags that can occur with multiple software systems and processes. That, in turn, makes creating a new decision support system easier, because there is less custom development involved.
Although all of these features aren’t available as a comprehensive decision support tool today, Flyzik expects that to change.
“Today you can create a fairly robust solution, but we are three to five years away from where we can take it for granted,” he said. Enablers over the next few years include increased bandwidth and use of the cloud, which will enable more end-to-end, holistic approaches and solutions. That’s because it helps reduce some of the infrastructure complexity by consolidating and managing it more as a centralized resource as opposed to many small different pieces, Flyzik added.