> The lack of data and analysis in this week's presentations was extremely 
> disturbing.

In considering protocol solutions to performance problems it is often very 
difficult to come to consensus because:

a. Performance data is often highly dependent on the implementations being 
studied, and therefore it is often unclear whether the conclusions hold 
universally. 

b. There are often multiple solutions available to the problem, some of 
which will perform better than others in given use cases.  Without 
agreement on the use cases, it can therefore be difficult to agree which 
solutions are superior, even if the performance data itself is 
unassailable (which is itself rare). 

The end result of this is that in practice, standards bodies often 
face the choice of waiting for completion of a protracted 
requirements/performance test development phase, or allowing work to 
proceed without a solid basis, under the theory that "the market will 
sort it out".  While the latter approach frequently results in solutions 
in search of a problem, this may at times be preferrable to deadlock. 

_______________________________________________
Int-area mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/int-area

Reply via email to