Alex Rousskov wrote:
If channel establishment is slow (because of XML or whatever), we have
a problem. The only way to solve this performance problem is to reuse
channels across OCP transactions. This is usually possible from BEEP
point of view (AFAIK), but makes the protocol more difficult to
understand and implement because you lose nice encapsulation of
channels within OCP transaction.
Did we convince ourselves that this kind of XML usage is a serious
(performance) problem? So far, we mostly speculated on the overhead
and possible performance impact without having hard facts. Does anyone
have concrete numbers or measurments to share?
If we're not convinced that this is a big, serious problem, we might
want to move on and focus our time and effort on getting OCP done
(rather than defining a transport protocol).