Posted on by Acquisitionin
By Nickolas H. Guertin Senior Software Systems Engineer Software Solutions Division
This post is also co-authored by Douglas C. Schmidt and William Scherlis.
In its effort to increase the capability of the warfighter, the Department of Defense (DoD) has made incremental changes in its acquisition practices for building and deploying military capacity. This capacity can be viewed as "platforms" (tanks, ships, aircraft, etc.) and the mission system "payloads" (sensors, command and control, weapons, etc.) that are populated onto those platforms to deliver the desired capability. This blog post, the first in a series excerpted from a recently published paper, explores opportunities in modularity and open systems architectures with the aim of helping the DoD deliver higher quality software to the warfighter with far greater innovation in less time.
The DoD cannot improve the capability of its existing portfolio of systems by relying solely on evolutionary upgrades achieved through its standard practices. Many perceive the evolutionary path as less risky, less time consuming, and less costly than instantiating a new product. Many existing products, however, were not designed to support incremental upgrades or even routine ongoing software and hardware sustainment. Instead, those systems were purpose-built and thus not structured to rapidly integrate new performance or scale to address adjacent solution opportunities.
The weight of current DoD architectures slows progress and increases risk to operational forces. What's more, rejecting 21st century development methods actually increases program risk and extends development timelines, effectively reducing the value of the DoD's acquisition portfolio. As a result, the current capacity for breadth and pace of change impedes our ability to evolve capability quickly and robustly enough to meet new requirements in emerging technical and warfighting environments.
There are now strong mission drivers to effectively exploit emerging technical capabilities such as machine learning, autonomy, and framework-deployed data analytics. Moreover, our adversaries are rapidly improving technical capability and demonstrating that they can be nimble in response to opportunities. This creates an engineering imperative to sustain sufficient capability and nimbleness to maintain a consistent technical lead.
This post is the first installment in a series examining how layered business and technical architectures can leverage modular component design practices to establish new approaches for capability acquisition that are more effective than existing system of systems (SoS) strategies. In this post, we will examine emerging opportunities in modularity and open systems architectures. Subsequent posts in this series will examine the following:
Emerging Trends and Opportunities
Addressing the limitations of conventional acquisition approaches requires a new set of business and technical practices to achieve different results and more advanced capacities than our adversaries. In particular, new acquisition structures and associated technical architectures are needed to harness the innovation engines of all sectors of the American and global economies. A key characteristic of applying a modular open system approach (MOSA) is that different components can be created by independent parties and can evolve at different rates.
When the DoD relies on the ecosystem that makes MOSA attractive, it loses some control but gains by "riding the growth curves" of capability and quality for commercial and open source technical ecosystems. Consequently, conventional approaches must be rethought at every level, including the ways that the DoD
Examples of Modular Open System Approach Adoption in the DoD
Segments of the DoD have aggressively innovated their acquisition practices. In each case, there was a driving need to change how a capability would be developed and delivered to include innovations in financial and programmatic alignment. Here are several representative examples of MOSA adoption within the DoD:
In general, industry has adopted these MOSA initiatives for three primary reasons: (1) to avoid being left behind as others find new opportunities and (2) to take advantage of new methods to improve internal corporate efficiency, as well as to (3) increase market share and increase profits.
Trends and Opportunities Enabled by Advances in Technology and Strategy
Changes in underlying COTS technologies used by the MOSA-enabled DoD programs described earlier in this post continue to evolve due to innovations in software technologies and architectures. It is now feasible to address backward compatibility and to use a variety of hardware implementations in any one system instantiation or data center while new technologies continue to evolve. For example, broad adoption of software patterns (such as Wrapper Façade) and host infrastructure middleware (such as virtual machines, virtualization, and containerization) enable greater support for backward compatibility of software onto other operating system and hardware environments by invoking COTS standard infrastructure (such as using graphic processing units to perform artificial intelligence processing in commodity cloud environments).
The COTS software building blocks available to develop, deliver, and manage capability have matured in commercial markets. It is now time to take a fresh look at how the acquisition, testing, and resourcing communities are structured to develop and rapidly deliver highly reliable, intuitively operable innovation in warfighting capability. One enabling step recently taken by the Navy was the establishment of the Digital Warfare Office (DWO), which is a leader in the area of decomposing the performance attributes of a system into functions.
The DWO focuses on methods for decomposing capability into elements that are internally tightly coupled but loosely coupled externally. These elements can then be applied to illuminate software modules needed to deliver the required military performance. It would be tempting to stop there and create a specification for a system that would comprise these functional elements. To reach greater performance and speed capability delivery, however, the Navy (and more broadly the DoD) must then extend this logic to structure the technical architecture to facilitate continuous delivery of innovations and avoid current independent system-based delivery epochs, which classically stretch from two to five years.
An Architecture-First Approach
A new architecture-first strategy is needed that embraces diversity and creativity, addresses enterprise performance equities, establishes conformant quality attributes (such as dependability, reusability, and flexibility), and manages variability while sustaining minimally coupled and inherently interoperable designs. This strategy should establish a framework of support infrastructure that provides an integration environment in which modules of capability can be hosted. New development methods and architectural constructs facilitate loose coupling of capabilities and deployment of software onto containerized or virtualized environments, thereby eliminating the need for hardware-dependent deployments.
Establishing and codifying rules of construction is the primary function of an architecture-first strategy that will facilitate creative alternatives. These rules are set to ensure quality attributes are known and followed throughout the lifecycle of a warfighting system. Likewise, these rules also ensure that warfighting systems support loose coupling (which enhances systematic reuse), low cyclicality (which is a metric that illuminates corruption of the benefits of modularity through over-indulgent interplay across modules), and quality attributes (often referred to as non-functional requirements).
While components and functions are separated, it is nonetheless the case that mission capability can be manifested to operators in a tightly-integrated manner. This integration is a consequence of effectively "matrixing" component capabilities and the design of both end-user experience and the application program interfaces found in traditional SoS.
Trends and Opportunities Enabled by Advances in Hardware
Advances in COTS hardware--coupled with an emphasis on software architectures that can readily adapt to hardware evolution--are enabling new opportunities for a hardware support model that facilitates continuous deployment of warfighting capability. The past practice has been to configure a specific hardware baseline, procure precisely those parts as a block-buy that will last the life of the deployed configuration (anywhere from 10 to 20 years), and then plan for the program to not run out of spare parts. Block-upgrades, however, are not a sustainable business model for either the commercial sector or the DoD.
Innovations in hardware sustainment strategies have fundamentally changed the methods and mechanisms of retaining the high-end capability needed by any organization whose business depends on modern data centers and cloud computing environments. These technologies support advanced software-centric development frameworks, such as virtualization and modularization. These development frameworks anticipate the rapid evolution of computing infrastructure, rather than locking systems into obsolescing or non-mainstream proprietary architectures. For example, commercial organizations, such as Google, Amazon, LinkedIn, and Facebook, apply these technologies to upgrade their data centers continuously with new hardware in a manner that allows them to deploy new software capabilities rapidly and reliably.
Forerunners of Advances in Acquisition Models
In recent years, various DoD efforts have combined several programs-of-record (PoR) to improve efficiency and to commonly do what is commonly done. For example, the SubLAN architecture from 2004 was a progenitor of the broader Naval effort for providing Infrastructure as a Service (IaaS) and Platform as a Service (PaaS) capacity to host non-tactical and crew services capabilities under the auspices of the Consolidated Afloat Network Enterprise Services (CANES). As mentioned earlier, CANES is consolidating and modernizing shipboard, submarine, and shore-based command, control, communications, computers and intelligence (C4I) networked systems to increase capability and affordability.
All Transformation Efforts Must Address Culture
The largest challenges faced by MOSA-related efforts are generally programmatic and cultural. For example, the enterprise-focused transformation effort of CANES provided PEO C4I with a common compute-plant and integration platform designed to meet the infrastructure needs of the Navy's C4I suite. It was relatively straightforward for the CANES architects to articulate the value proposition of integrating common artifacts and components that were not initially designed for common use.
However, PEO C4I also established other programmatic elements, including a shared and evolving build environment, the capacity to host a wide array of capabilities, and a PEO C4I organizational policy of rewarding the creation of common elements. Obtaining buy-in for these elements was therefore harder since they were antithetical to the classic acquisition behavior of protecting the PoR and preferring a system-by-system, go-it-alone approach. In general, PEOs will need to face these types of portfolio optimization issues directly if they wish to pursue a completely redesigned approach for continuous delivery of modularized, advanced, reliable, and innovative capability into a continuously modernized and shared environment.
The Defense Acquisition Executive (formerly USD [AT&L]) has recently been split into the undersecretaries of Research & Engineering (R&E) and Acquisition & Sustainment (A&S). This decomposition shows a path for organizing around the principles of focusing on innovation for the warfighting domain (the R&E portfolio). It also shows a path for devising a highly reliable and flexible integration environment for those innovations (the responsibilities of A&S).
One of the most valuable outcomes of splitting these activities is the acknowledgment that each entity works on different activities, with different skillsets and business drivers, yet each must depend on the other if either is to succeed. The organizational construct of the former USD (AT&L) was predicated on a different strategy and orientation of engagement for oversight of acquisitions performed by the military services (Army, Navy, Air Force and the "Fourth Estate"). The existing staff will have to undergo a deep culture change if the split into R&E and A&S is to succeed.
Note, however, that a reorganization alone won't solve the problem of how system and software architectures can be jointly owned and managed by government and (multiple) contractors. What is also necessary, therefore, is for the DoD to invest in and develop sufficient government-aligned capabilities to make good decisions pre-RFP regarding both architecture and iterative increments/milestones, such as those emerging with Agile-at-scale methods. There is also a need to focus on methods for sustaining architectures over long system lifecycles. While companies like Apple and Google continually evolve their architectural frameworks (despite more than a million apps in each), the DoD programs have historically struggled to refactor their architectures in a cost-effective and timely manner.
Looking Ahead: Toward a New Model of Acquisition for the DoD
It is widely recognized that the DoD needs a nimble response to nimble adversaries, including incremental improvement to existing capabilities, granular delivery of new payloads, and the ability to continuously deliver to the military platform. The current pattern of upgrading ships and aircraft applies a system-by-system, "rip out and (re)install" process. This pattern, however, incurs prolonged periods to upgrade capabilities, reduces operational availability, and impedes interoperability.
Another area that is widely agreed to in principle--but has been even more elusive to achieve in practice--is taking successful prototypes and productionizing the capability with excellent quality, full support, and training. The benefit of rapidly attempting new ideas and quickly declaring success or failure may be lost, however, if those prototypes are fielded in a way that does not match the business needs of the organization. Without good architecture practices, those efforts might provide a near-term salve on an urgent problem.
Advances in COTS hardware are enabling new opportunities for a hardware support model that facilitates continuous deployment of warfighting capabilities. Our next post in this series will examine a new approach that applies a coherent set of methods to develop military mission capabilities as sets of composed modules. Future posts will present key change drivers and technical and/or organization structures associated with the new model of acquisition we propose for the DoD. Implementing these recommendations will establish a DoD acquisition environment shaped to be more efficient, deliver much higher quality--with far greater innovation--in a fraction of the time.
This post was excerpted from a recently published paper, Capability Composition and Data Interoperability to Achieve More Effective Results than DoD System-of-Systems Strategies, which examines a new model of acquisition that applies a coherent set of methods to develop military mission capabilities as sets of composed modules.
Read the SEI Blog Post by Don Firesmith Open Systems Architectures: When and Where to Be Closed.