Go open to transform legacy systems
- By Adam Firestone
- Dec 18, 2013
Many legacy military, intelligence and other government systems display software architecture patterns, including monolithic construction and proprietary interfaces, that make sustainment unacceptably resource-intensive. Nevertheless, those systems are slated for retention because:
- They continue to fulfill operational requirements.
- They are globally deployed on many platforms with idiosyncratic maintenance schedules.
- Development is done by multiple contractors.
- The cost of updating the systems approaches or exceeds sustainment costs.
Solutions based on a service-oriented architecture (SOA) enable the transformation of legacy capabilities into systems that are modern, modular and sustainable. To date, achievable implementations have been frustrated by the lack of suitable software platforms, the large acquisition costs of such platforms, a shortage of effective life cycle governance mechanisms, or a combination of all three.
Now powerful open-source platforms are available to address those concerns.
In a SOA, "separation of concerns" is a design principle requiring a system's separation into sections, each addressing a discrete functionality. The approach simplifies and economizes maintenance. System elements can be developed and maintained independently, and operations can be conducted on one section without having an impact on others.
From the perspective of an SOA architect, the most fundamental separation of concerns is the one between domain capabilities (i.e., business or warfighting processes) and infrastructure capabilities.
Issues created by monolithic architectures and proprietary interfaces require an architectural solution that separates domain from infrastructure concerns, eliminates redundant capability, is readily adaptable and can interoperate with legacy systems.
Open-source software offers solutions to the technical, budgetary and governance challenges that have frustrated efforts at transformation.
The solution's core construct is a generic enterprise integration platform. The EIP provides system infrastructure capabilities that support domain-specific applications and services and offer the means for integrating legacy components.
Build or buy is no longer the question; commercial packages are the only realistic alternative. Unfortunately, many of them come with licensing costs sufficient to offset sustainment savings associated with the objective system. Fortunately, a number of open-source packages are addressing the issues of functionality and cost.
Open-source software is supported by government guidance. For example, a Defense Department policy from 2009 declares open source to be equivalent to proprietary software. The policy instructs the acquisition community to evaluate open-source software comparably to any other software. As a result, it is now feasible to design an EIP around open-source components. And adherence to open standards alleviates integration difficulties.
Such an EIP is a deployable system core that avoids licensing costs. It provides infrastructure capabilities that support standards-based domain services. It also allows legacy components to be incorporated through the use of integration technologies, such as an enterprise service bus.
Transformation includes modernizing domain-specific capabilities. That challenge is less about what should be produced and more about how production should be managed across multiple developers, time zones, technical postures and management approaches. An effective solution ensures that processes are common across the program and controlled by the government program manager.
Fortunately, there are cloud-based development toolkits for open-source software that automate programmatic governance. They provide self-service provisioning and continuous build, integration, test and delivery, resulting in increased developer productivity, shorter project cycle times and more efficient resource use. The program manager benefits from improved programmatic predictability, risk reduction and lower program cost. As with the EIP, this space contains a number of open-source offerings.
In sum, open-source software offers solutions to the technical, budgetary and governance challenges that have frustrated efforts at transformation. Perhaps the most intriguing question about legacy system transformation is not whether it can happen but why it hasn't happened yet.