DO-178C and DO-254: Why Aerospace Certification Is Becoming Unworkable Without a Connected Baseline
Aerospace certification is the most expensive engineering activity on Earth per kilogram of delivered product. A 2025 RTCA benchmark put the average DO-178C Level A software certification at €18 to €25 million per program, and DO-254 hardware certification for complex ASICs and FPGAs at €8 to €15 million. These numbers have roughly doubled since 2015, while the underlying engineering content has not. The cost driver is traceability evidence, not technology.
The Evidence Problem
DO-178C requires bidirectional traceability between system requirements, software requirements, design, source code, and test cases. DO-254 requires the same for hardware items down to the gate level. The standards were written assuming a structured, reviewable artifact chain. In practice, most programs deliver this chain as a collection of Word documents, Excel matrices, and DOORS exports stitched together in the final months before SOI-4.
The EASA published a 2025 oversight report flagging traceability gaps as the leading cause of certification delays on new programs, ahead of test failures and design rework (EASA Annual Safety Review 2025, section 4.2). The gap is rarely a missing artifact. It is the inability to prove, on demand, that a given line of code or a given FPGA block traces back to a derived requirement that traces back to a system safety objective.
What Has Changed in 2026
Two regulatory shifts have made the manual approach untenable. First, EASA and FAA both updated their guidance in late 2025 to require continuous traceability evidence for software updates under field-loadable software provisions, not just initial certification. Second, the rise of complex AI-enabled avionics under the EASA AI Roadmap (Concept Paper Issue 02, 2025) introduces traceability requirements for training data, model architecture, and runtime monitoring that no document-based system can satisfy.
The defense side adds another layer. The French DGA and the US DoD have both moved to require Model-Based Systems Engineering deliverables on new contracts above €50 million as of 2026 (DGA Instruction 1516, January 2026). MBSE without a connected backend is just SysML diagrams in a vault. With one, it becomes the certification evidence itself.
The Cost of Reconstructing Context
The hidden cost in aerospace programs is not the certification audit itself. It is the engineering time spent reconstructing context for every formal review. Engineers on Level A programs report spending 30 to 40 percent of their time on traceability artifacts rather than on engineering work (NASA TM-2024-001, internal productivity benchmark). On a 200-engineer program over five years, that is roughly 350 person-years of avoidable work.
A connected baseline that holds requirements, architecture, code, hardware items, and test evidence as queryable nodes turns this overhead into a query. Stage of Involvement reviews stop being multi-month preparation efforts and start being a status report against the live system.
What Aerospace Programs Should Do
The first move is to stop treating traceability as a deliverable produced at the end of the V-cycle. It must be a live property of the engineering data itself. The second move is to consolidate the artifact chain into a single graph where every link is enforceable and queryable. The third move is to align the safety case, the certification evidence, and the engineering data on the same source of truth.
The Koddex Position
Koddex holds the full V-cycle as a connected graph: system requirements, software and hardware items, design artifacts, verification evidence, and safety objectives. DO-178C and DO-254 traceability stops being a quarterly fire drill and becomes a live capability. Aerospace programs running on Koddex spend their certification budget on engineering quality, not on rebuilding evidence chains under deadline pressure.
