On July 25, 2008, the Honorable John Young, Under Secretary of Defense for Acquisition, Technology and Logistics (AT&L), signed a memorandum for service acquisition executives that represents a fundamental reengineering of acquisition management systems. The memo "Implementation of Service Oriented Architecture (SOA) within DoD Acquisition Community," aims to significantly improve accurate, timely and coherent data use within the acquisition community to enable better decision making. How we got to this decision and what it means for Defense Department managers and the military services is the subject of this article. The story is remarkable, not only because of its success, but the speed at which it was executed.
Figure 1 represents the data model currently used by the DoD to make acquisition decisions — and a proposed model — a model many businesses have used since the 1980s. Actually, Figure 1 fails to highlight one of the most troublesome flaws of the current system: data is provided via the conduits indicated on an extraordinarily slow basis. At each level, data is reviewed and forwarded to the next level but may no longer be accurate by the time it is reviewed at the next level.
One of the most desirable features of the new DoD model is that each request for data from the indicated SOA server generates a separate call from authoritative sources for the data element. In this concept called "data as a service," a data request invokes an action that immediately returns the authoritative data element with a clear origin.
The question is: How do we move DoD's many and varied systems to a model that separates data from these systems for a streamlined approach to data management?
Feb. 29, 2008, marked the start of a change in the way DoD manages defense acquisition data. On that day, three business tools displayed authoritative data from 12 Major Defense Acquisition Programs on demand for the Weapon Systems Lifecycle Management (WSLM) Core Business Mission (CBM) senior steering group and invited DoD executive leadership.
Data for 12 MDAPs, four from each service, were obtained from the authoritative source for each data element in real time. To further prove the flexibility of this new data management model, the Air Force changed one of its authoritative data elements at the program management source for the B-2-EHF SATCOM program. Business tool displays were refreshed, and the revised program data element was properly reflected in each display and available for use by DoD in less than two minutes.
This demonstration of data governance and technical capability offers acquisition officials the ability to make informed acquisition decisions based on timely and authoritative data.
USD (AT&L) initiated this effort with a memorandum issued Oct. 5, 2007, mandating a Service Oriented Architecture (SOA) Demonstration Project (AT&L SOA Demo). The demo, which later successfully expanded to a pilot project, was co-led by Gary R. Bliss, Deputy Director of Enterprise Information and OSD Studies in the office of Acquisition Resource and Analysis (ARA), and Mark E. Krzysko, Assistant Deputy Under Secretary of Defense for Business Transformation (AT&L(BT)), both within AT&L, and with participation by the military services and the Business Transformation Agency.
Providing strategic functional vision and leadership, Bliss established the foundational data governance and management concepts and methodologies to be employed and Krzysko provided technical vision and leadership, establishing the SOA middleware framework and SOA infrastructure required for the demonstration.
AT&L ARA staff developed functional requirements and the policy framework. The Business Transformation Office published a straightforward technical approach and arranged to host the SOA infrastructure.
In the proceeding months, Bliss and Krzysko led all aspects of the AT&L SOA demonstration to triumphant completion.
SOA is not just an architectural style of services seen from a technology perspective, it encompasses the policies, practices and business processes through which governance ensures the right information is provided to, and consumed by, users.
SOA separates data governance from the tools that use the data, thus making authoritative data immediately available to users. Under SOA, data is pulled directly from sources the military services and other data providers designate as “authoritative.” These data elements are then “federated” and made available for consump-tion by any authorized user through any authorized tool.
A federated database, or virtual database system, is a type of meta-database management system which transparently integrates multiple autonomous database systems into a single federated or merged database. The databases are interconnected via a computer network and can be geographically decentralized.
The problem that needs to be overcome is not simple. For the most part, current data systems were implemented to electronically mimic pre-existing paper-based systems. SOA is reaching for a profoundly different data model; one in which data is transparently available throughout the enterprise to whomever has a legitimate need for the data as soon as it is developed.
The existence of a well-functioning SOA infrastructure does not eliminate periodic program reviews and the data displays associated with them. On the contrary, it profoundly transforms program reviews. Instead of just providing raw data, data will be interpreted for decision making by senior acquisition managers. This change is immense!
SOA Governance and Data Elements
SOA requires a SOA governance institution. The creation of the governance structure during the AT&L SOA Demo, shown in Figure 2, among others, includes acquisition officials from AT&L(ARA), AT&L(BT), Army, Navy, Air Force and Defense Acquisition Management Information Retrieval.
DAMIR is a DoD system that provides enterprise visibility to acquisition program information. The primary goal of DAMIR is to streamline acquisition management and oversight by leveraging the capabilities of a net-centric environment.
In its current configuration, the tool must encompass all the means for obtaining and purifying its own data. Even so, maintaining data coherence with sources is a perpetual — and losing strug¬gle. Under SOA, DAMIR is re-plumbed to extract data from the SOA server, which is already authoritative. The difference for DAMIR will be immense (almost no data effort), and perhaps new tools of unanticipated sophistication and functionality can be easily plugged into the SOA data server, and all the tools will be using the same data at any point in time.
The WSLM CBM data governance structure has a charter to address the three key elements of data infrastructure management: data definitions, technical standards and unambiguous assignment of institutional responsibility to maintain the single authoritative copy of each governed data element.
The SOA Demonstration leveraged 61 data elements associated with the management of 12 MDAPs. The results showed that commercial information technology tools, with careful regulation of the definition and technical standards, permit secure and transparent use of data from disparate sources that facilitate an acquisition user-defined operating picture (UDOP) for business systems, similar to use in warfighting systems.
Data brought under governance for the subsequent pilot include 140 elements in the following major categories:
• EV (Earned Value) Data – Elements used in the demo plus additional contract elements included in DAMIR’s “Contract Data Point” and/or reported on the Contract Performance Report (CPR).
• Nunn-McCurdy Unit Cost – Current Estimate, Current Acquisition Program Baseline (APB) and the original APB Unit Cost Data reported at the total appropriation level (i.e., RDT&E, Procurement, MILCON and O&M).
• Budget Submission – Last president’s budget and Program Objective Memorandum (POM)/Budget Estimate Submission (BES) by appropriation and fiscal year to provide a reference point for POM 10 analysis.
• Milestone – Program acquisition milestones as agreed from APB(s).
• Science and Technology – Key Performance Parameters, thresholds, objectives and current measurement.
• Program Management (General) – General program administration elements.
Through this governance body, agreement to data definitions, identification of authoritative sources for the required data, and achievement of data availability for on-demand access were accomplished for the demonstration and pilot. A central management focus is required to coordinate business processes and information requirements to avoid duplicative efforts and costs while ¬encouraging the reuse of information services to promote more efficient acquisition management decisions. Cost avoidance will present a return on investment. Other potential long-term benefits of SOA are:
• Establishing managed authoritative DoD acquisition data sources;
• Improving data availability and reliability to decision makers;
• Improving situational awareness of the acquisition status of each MDAP;
• Separating data from business tools and applications;
• Improving program management and oversight efficiencies;
• Reducing burdensome oversight reporting; and
• Reducing the acquisition cost for future business systems.
A critically important governance function is to identify where responsibility for maintaining the single authoritative copy of an individual data element should reside. For those data elements for which the service acquisition executives and program executive offices are responsible, there is no problem with the services controlling their own data.
The vast majority of data for a program, however, is what may be called “state data” such as the program name, fiscal year, department code and contract number. Under SOA, these data would be transparently visible throughout the enterprise.
For users, there is one place to go for any data element governed by the SOA middleware server. Each data element would come with date and time stamps: the “pull date” when the data was accessed and the “shelf date” when the data was last updated.
Data maintainers will have a streamlined method of updating data. The SOA governance mechanism will issue documents that explicitly identify what offices are responsible for maintaining each individual data element. In general, data maintainers will be able to ensure that a specific spreadsheet, for example, is kept current with the correct values and stored in a specific server directory. The result will be the elimination of continuous management requests for updated data tables.
SOA Pilot Project
Based on the success and lessons learned from the AT&L SOA Demo, a two-phased pilot project is underway led by Bliss and Krzysko. Phase 1 was initiated in April 2008, and the initial capabilities will be completed by fall 2008. An additional 25 MDAPs have been added to the original 12 for a total of 37 which equates to acquisition data representing 75 percent of existing MDAPs — a $1.3 trillion portfolio.
The technical center for the SOA Acquisition Visibility pilots was established at the Space and Naval Warfare Systems Center Atlantic, formerly SSC Charleston, with David Howard as the infrastructure lead, supporting Bliss and Krzysko.
Through Phase 1 efforts, defense acquisition decision making will move to a new paradigm of “seamless authoritative data transparency” based upon underlying governance provided by the WSLM. Figure 3 illustrates the technical approach.
Phase 2, performed concurrently with Phase 1, involves developing an initial operating capability for a DoD-wide SOA infrastructure compliant with pertinent DoD IT standards to include full competitive source selection for technical products or solutions.
The primary goal of this effort is to make ACAT I programs data consistent across the DoD. As this is accomplished, the WSLM CBM, which includes representation from other communities, such as systems engineering, science and technology, test and evaluation and contract management, will work to align data using WSLM definitions. The result will be the establishment of a durable data governance mechanism for regulating DoD’s acquisition data, as well as the provision of a data infrastructure and SOA services, which makes data as a service a reality in the acquisition domain.
Martin Fairclough is a senior principal consultant supporting the AT&L SOA Team.