Mapping architectures to the Viable System Model

Been thinking of another way to distinguish between enterprise architecture (the real ‘whole of enterprise’ form, not IT-architecture masquerading as ‘enterprise’ architecture) and business architecture, because at first glance they seem to cover much the same scope.

Perhaps the simplest way to show the differences is to map them to service-oriented architecture and thence to Stafford Beer’s Viable System Model . The VSM consists of a recursive set of six ‘systems’:

• the ‘management-services’ set:
  • system-5: policy and purpose – “decisions to maintain identity”
  • system-4: strategy, ‘outside, future’ – “development, research and marketing”
  • system-3: tactics, ‘inside, now’ – “operations planning and control”
• the ‘pervasive-services’ set:
  • system-3*: audit and quality-management
• the coordination-services set:
  • system-2: coordination for ‘develop the business’, ‘change the business’, ‘run the business’ – “regulation and tactical planning”
• the ‘delivery-services’ set:
  • system-1: subsidiary services or tasks to enact the actual service-delivery (each of which is also a Viable System with its own system-5, -4, -3 etc)

Almost by definition, architectures all tend to operate somewhere as or between system-3* and system-2 – representing and promulgating enterprise-values such as consistency, integration and effectiveness, and assisting in coordination to enhance these. But in terms of scope:

• enterprise architecture deals with system-5 type concerns, bridging over into system-4 and on ‘downward’ from there
• business-architecture has its emphasis on system-4, with some cross-links into system-3
• TOGAF’s ‘information-systems architectures’ tend to sit in system-3, with some links into system-2, down into system-1, and in some parts into system-3*
• process-architectures and TOGAF’s ‘technology infrastructure architecture’ sit fairly solidly in system-1

Seems to make more sense in those terms, anyway.