Hitting the software wall: the limits of consolidation in ADAS and automated driving systems

The reason ADAS and automated driving resist consolidation is not that safety-critical software is harder to write; it is that the consequences of getting it wrong are categorically different. Regulators are agnostic on whether a function runs on shared or separated hardware, but they require evidence that safety performance holds regardless of architecture, and that evidence is more expensive to produce on a consolidated platform.

Hardware consolidation in modern vehicles is no longer the industry's defining challenge, although it remains far from complete. The direction of travel is clear: from dozens of discrete electronic control units (ECUs) toward fewer, more capable compute platforms. Domain-centric solutions, including ADAS domain controllers (ADCs) and cockpit domain controllers (CDCs), have been the dominant intermediate step and are likely to remain the largest part of the industry for at least the next decade. Beyond domain controllers, zonal and central compute architectures represent the more ambitious end of the spectrum, promising further reductions in wiring-harness mass, procurement complexity, and capex on connectors, brackets, and bespoke microcontroller units (MCUs). This part of the software-defined vehicle transition is, at this point, well-rehearsed and broadly agreed, although not yet realized in any significant share of the market, in large part because of the mixed-criticality software problems these architectures inherit.