“Our adversaries modify TTPs (Tactics, Techniques, and Procedures) and adopt new technology faster than our acquisition system can respond.” In a world with increasing chaos and declining budgets, this seems to be a perennial sentiment: “We need speed and agility, pulling capabilities off the shelf, delivering capabilities in months, not years.”
Sound familiar? Unfortunately, the conversation usually ends there…a tacit admission of defeat lacking any intention to correct the problem. Rather than being the end of the conversation, it should be the beginning.
Some root cause assessment is in order, the question being what makes the situation so? What takes the most time and resources? The answer:
1) research and software development time to solve hard problems at scale
2) time for integration (and re-integration) of IT infrastructure
These “twins” of traditional cost plus systems acquisitions have the unintended consequence of incentivizing industry to pursue head-count based “services” revenue rather than “capability” revenue. Integrating systems over and over is good for business when head count is a convenient proxy for “capability” delivered. Rather than adopting existing capability or buying available off-the-shelf software, programs opt to develop program-specific software and an IT infrastructure optimized to mask development flaws affecting system performance.
Take, for example, analyst tools. At a basic level, analytic timelines are dependent on the time it takes to assemble data, read and assess the material. The DoD and IC have been pursuing automation in this area for years using a “system” approach. Such a system would ingest data, read material, and provide analysts with context assertions, relationships, graphs, and other unstructured data to derive building blocks for Activity Based Intelligence. The traditional procurement involves requirements definition, followed by functional decomposition to an atomic level (including defining the IT infrastructure), developing software in a program-dedicated development environment, integrating multiple software and hardware components, testing and accrediting for high side use…and then re-testing because (pick a problem)…you get the picture.
THIS IS WHERE OUR ANALYSTS BECOME THE VICTIMS OF AN UNRESPONSIVE ACQUISITION SYSTEM…TO OUR ADVERSARIES’ ADVANTAGE.
To be fair, these are difficult problems. High accuracy unstructured data analytics at scale with acceptable timelines and acceptable false alarm rates is difficult business. Then again, why “over-complexify” the problem by starting from scratch on the software and surrounding infrastructure? Why embark on a 2-3 year in-house software development effort when the capability can be procured off the shelf? It’s not cheaper…life cycle cost and speed to mission for GOTS vs. COTS is a well-researched topic and the jury is out in favor of COTS.
Traditional systems development and integration models simply don’t close on cost or delivery timelines. Leveling the playing field with our adversaries requires a different approach, one the DoD and IC are embracing in pockets despite the institutional inertia against it. They are moving away from vertically integrated systems toward cloud services, specifically pre-integrated “platforms”. These cloud based platforms address the speed to mission challenge by hitting the two “long poles” of the acquisition system head on:
- Pre-integration of IT infrastructure environments with provisioning of platforms for development and production.
- Integrating “ensembles” of analytic software, combining off the shelf (COTS) with government (GOTS) technology where necessary to address specific, niche requirements unmet by existing COTS products.
This move, beyond procurement of integrated systems to procurement of products, integrated software, and provisioned IT infrastructure environments, brought together into mission threads by an integrating organization specifically prohibited from developing the components, is overdue. Questions remain, principally, what is the appropriate level of “pre-integration”? We know that our traditional integration of hundreds of components isn’t conducive to speed to mission. But it’s clear that integration of fewer, larger components can dramatically reduce the complexity and cost of integration.
Proactive companies have recognized the trend, emerging as software integrators of pre-integrated “building blocks” of functionality and components. These will be the new market leaders. Tenets they live by:
- “ABC: Adopt, Buy, Create” to minimize cost of goods sold. Only develop that to which there is no existing alternative…and be intellectually honest in that determination by considering product roadmaps.
- Computing performance has progressed to a point where custom tuned systems are unnecessary for most applications. Anything I need to roll on my own is a margin-killer.
- Open API’s and Software Development Kits are table stakes. You don’t get to play if you’re unwilling to allow others to leverage your capability and extend your functionality with their ideas.
One of the IC’s leading analytics firms, Digital Reasoning, is leading this wave of change with the upcoming integration of its Synthesys® platform running on the Amazon’s government and secure cloud platforms. Synthesys® is a pre-integrated ensemble of entity, relationship, and advanced analytics models that in turn integrates with multiple third party data sources and visualization engines to ensure the right method is used on the right data to deliver the most accurate and contextually relevant information. Using the C2S “Platform” as its pre-integrated IT environment, Digital Reasoning develops new capabilities on the fly using unclassified AWS instances and drops them into the classified C2S environment in a matter of weeks, not months or years. As a result, analysts from across the IC are able to automate the process of reading, resolving, and reasoning on the myriad of intelligence data used in their analysis and assessment work. More importantly, the software is available to use in other mission applications without retooling or re-hosting. Rather than spending precious development resources (and time we don’t have) building Activity Based Intelligence and Object Based Production systems from scratch, the IC is able to leverage existing (paid for) capabilities and, in fact, rapidly extend them via the provisioned IT platform.
Tim Estes, CEO of Digital Reasoning, puts it this way, “Pre-integration of analytics, combined with cloud platform environments, has the potential to minimize development risk, deliver critical capabilities we don’t have today, reduce cost and time to mission – while outpacing the adversary.”
As the old adage goes, doing the same thing and expecting different results is folly.
Maybe we need to get serious about this new way of doing business.
This post and related information can also be found by visiting the AnalystOne site.
About the author:
Captain Buck completed a 26-year Navy career with over 15 years in National Security Space, including 10 years in the National Reconnaissance Office. A Naval Aviator and Test Pilot with over 2,500 flight hours, Nick has served as both a supplier and a customer of Intelligence, Surveillance and Reconnaissance Information Technology systems. Nick managed multiple major defense acquisition programs and is a recognized expert in implementing Information Technology consolidation and Cloud Computing programs. A Harvard Senior Executive Fellow, Nick currently serves as President and Chief Executive Officer at Buck Consulting Group, LLC.