Etienne Urbah proposed this requirement on 2010-03-17

Etienne Urbah's position on 2010-04-07

• SPOFs (Single Points of Failure) and bottlenecks are often caused by careless implementation, but are sometimes caused by lack of consideration of such problems during the specification phase.
  
• Therefore, specifications should :
  - NOT force or implicitly assume that services will be performed by a single server with failure-free unlimited power,
  - but ease implementation of services by redundant application servers, possibly driven by a load management server.
• This redundancy could be achieved for example by using dynamic DNS, stateless protocol, ...
• Without more precise recommendations, this requirement of me is just wishful thinking.

Morris, Balazs and Etienne on 2010-04-27

• Source = General
• Spreadsheet ID = 149

Amsterdam meeting on 2010-04-28

• Agreed YES