Significance Airborne electro-optical (EO) systems are key sensors at the front end of the kill chain, and their ability to iterate rapidly and upgrade at low cost directly affects the operational effectiveness of the platform. Traditional airborne EO systems, however, are mostly delivered as physically integrated turrets or pods, in which detectors, processors and display-control equipment are tightly coupled, interface protocols are proprietary, and functions are solidified in dedicated processing hardware. As a result, capability upgrades are slow, costly and locked to single vendors. To solve such problems of tightly coupled C4ISR architectures, poor equipment reusability, protracted development cycles and escalating costs, the US Department of Defense has pursued open systems architecture initiatives since the 1990s under the top-level guidance of the modular open systems approach (MOSA). Among the resulting standards, the sensor open systems architecture (SOSA) has become the authoritative open architecture specification for sensor systems. Although domestic research on SOSA has accumulated considerable results in standard interpretation, interface specification analysis and local application of VPX hardware platforms, most of it remains at the level of technical translation, and a systematic transition method tailored to China's airborne EO systems is still absent. A systematic analysis of SOSA's technical concept, architecture ecosystem and engineering methodology, together with a feasible open design method for China's airborne EO systems, is therefore of practical significance.

Progress The SOSA open architecture is systematically analyzed. SOSA holds a unique "standard of standards" position within the MOSA spectrum: it references existing standards such as FACE, OpenVPX/VITA and MORA, supplements sensor-specific content, and abstracts a sensor system into mission, functional and service layers with six functional module groups. A complete conformance and certification ecosystem has been established, including Conformance Sets, the x242 verification procedure, the x215 certification policy, independent Verification Authorities and certified test tools. Engineering practices such as the Agile Condor pod have demonstrated SOSA's capability to support rapid iteration and cross-platform adaptation, and the agile mission suite government reference architecture (AMS GRA) v14.0, released by the US Department of the Air Force in April 2026, further transforms openness requirements into numbered, tiered and verifiable compliance provisions. Based on publicly reported case data, the quantitative benefits of open architectures over traditional closed architectures were compared in terms of integration cycle, software reuse rate and life-cycle cost, with the caliber of each figure (empirical value, engineering measurement, demonstration condition, statistical value or upper bound) explicitly annotated. The comparison shows that the integration cycle of a new software block has been reduced from an empirical 18-33 months to about six weeks in open mission system demonstrations, that software code reuse of up to 90% has been achieved among helicopter platforms of the same family, and that operation and support costs account for approximately 70% of the total life-cycle cost, where open architectures mainly deliver benefits through module-level replacement and technology insertion. Systematic implications for China's airborne EO systems are then distilled. At the methodological level, the function-logical-physical (FLP) three-domain separation methodology is refined: the functional domain is defined openly by the user or the requirements-authoring organization, the logical domain is defined by developers under unified guidance, and the physical domain is independently designed by developers under logical-domain constraints, so that the three domains can evolve independently at different paces. The EO working thread is standardized into seven links, namely optical front-end collection, analog front-end conversion, digital front-end sampling, signal processing, target processing, decision support and back-end processing, and its coupling and mapping relationship with the three domains is clarified through a dedicated block diagram with interface anchors I1-I6. At the engineering level, implications concerning open hardware architecture, service-oriented software architecture and plug-and-play mechanisms are elaborated with concrete application scenarios drawn from typical EO detection and tracking missions. At the mechanism level, a user-led, clause-numbered and assessable conformance governance mechanism is emphasized. Furthermore, a progressive open architecture (POA) design method is proposed. POA follows four principles: compatibility with existing assets, layered progression, controllable risk and user leadership. A three-stage implementation path is constructed: application-layer openness, which decouples algorithms from the platform and makes detection, recognition and tracking algorithms replaceable and competing; service-layer openness, which encapsulates platform functions as standardized services for cross-platform reuse; and hardware-layer openness, which restructures hardware into standardized module profiles and enables cross-vendor interchange and dual-source procurement. For each stage, objectives, key measures, risk fallback strategies, verification metrics and exit criteria are specified. Adaptability and risk analyses are conducted with respect to the differences between Chinese and US industrial systems and acquisition regimes, stringent airborne size, weight and power (SWaP) constraints, data-link and security requirements, and domestic substitution scenarios involving indigenous chips and operating systems, and a "profile unchanged, devices replaceable" substitution channel is established through device mapping.

Conclusions and Prospects The analysis confirms that SOSA, supported by its conformance ecosystem, has evolved from a technical standard into assessable compliance governance, and that open architectures have produced verifiable engineering evidence of order-of-magnitude reductions in integration cycle and substantial improvements in software reuse, while systematic life-cycle cost-benefit quantification remains an open issue. The FLP three-domain separation and the working-thread standardization provide methodological guidance for agile low-cost upgrades and for building an open supply-chain ecology, and the proposed POA method offers a smooth, controllable and stage-wise transition path from closed to open architectures for China's airborne EO systems. The conclusions provide a basis and reference for the development of China's airborne EO systems and for the demonstration of their open architecture roadmap. Future work will focus on functional-domain definition pilots, algorithm repository construction and verification environment development for POA, so as to promote the engineering application of the proposed method in equipment demonstration programs.