Considerable work has been done to extend the Constructive Systems Engineering Cost Model (COSYSMO) to estimate effort associated with the engineering of system of systems (SoS) capabilities (Lane 2009). This COSYSMO cost model extension for SoS is based on the SoS engineering (SoSE) processes described in the Department of Defense (DoD) Systems Engineering Guide for System of Systems (DoD 2008). Anecdotal evidence from many SoSE projects (NRC 1999; Sledge 2010) indicates that a major challenge in both maintaining existing SoS capabilities and developing new SoS capabilities is constituent system non-interoperability. For net-centric SoS, this interoperability issue is often related the compatibility of interface protocols and internal system data models that manage the data and information shared across the SoS constituent systems (Fisher 2006; Smith et al. 2006). This challenge is exacerbated as systems become part of multiple SoSs because there are inconsistent standards across SoSs.
The commercial sector has turned to Service-Oriented Architectures (SOAs) to deal with this problem for enterprise-wide solutions. This approach improves interoperability, but requires a major upfront investment and continual care and feeding as systems come and go within the SoS enterprise (Fisher 2006; Lewis et al. 2010) and is not always well-suited for certain environments such as applications where one way asynchronous communications are necessary, situations where loose coupling is not desired or even detrimental, or applications requiring an extensive graphical user interface with high levels of data exchange (e.g., map rendering and manipulation) (Exrforsys Inc. 2010). Even in those situations where SOAs may be appropriate, the organizations contributing to or “owning” the SoS may need to establish a SOA governance model to oversee SOA design decisions or guide extensive system modifications and refactoring for effective SOA implementation (IBM 2010).