"When it comes to procurement, for the better part of five decades, the trend has gone towards lower numbers as technology gains made each system more capable. In recent years these platforms have grown ever more baroque, ever more costly, are taking longer to build, and are being fielded in ever dwindling quantities."The US Department of Defense appears to have a very clear understanding of the opportunities for Service-Oriented Architecture in enhancing the procurement process and improving the economics of governance. In July 2008, the SOA Acquisition Working Group of the AFEI produced a useful set of Industry Recommendations for DoD Acquisition of Information Services and SOA Systems.
"The issue then becomes how we build this kind of innovative thinking and flexibility into our rigid procurement processes here at home. The key is to make sure that the strategy and risk assessment drives the procurement, rather than the other way around."
One point that comes across very clearly from this paper is the opportunity to separate the provision of different layers within a layered architecture. So we may have one set of suppliers providing the underlying layers, and a different set of suppliers providing the upper layers.
The AFEI paper identifies three roles - Platform Providers and Commodity Infrastructure, Component Developers and Enterprise Managers - with what it calls Firewalls between them. (See Figure 3 in the AFEI paper.) These firewalls roughly correspond to the first two of what Philip Boxer calls the three asymmetries of demand. The third asymmetry (not explicitly shown in the AFEI paper) is between the Enterprise Managers and the Customers. (Including customers in the schema is essential if you want to move towards a risk-reward model of procurement, as the DoD evidently does, because it is the customers who ultimately generate the effects.)
Managing all this complexity almost certainly results in higher management cost than if you simply outsource the whole lot to a large system integrator, but it also shifts the management control (architectural governance) back to the acquiring organization.
So which kind of organizations would want to do this then? It's a question of complexity - the greater the level of complexity and volatility in your requirements, the greater the potential benefits from having finer-grained and stratified control over your procurement.
And what kind of organizations are capable of this fine-grained stratified control? Perhaps not too many at the moment, which is why we need a capability maturity model for SOA that provides a roadmap for organizations to develop an appropriate level of SOA capability.
AFEI SOA Acquisition Working Group, "Industry Recommendations for DoD Acquisition of Information Services and SOA Systems", July 2008 (register on the AFEI website for free download)
For explanation of the three asymmetries, see Richard Veryard & Philip Boxer, Metropolis and SOA Governance, Microsoft Architecture Journal, July 2005.