J. Atwood wrote: <snip>

I do think, though, that performance and benchmarking should be an issues on the tips of DC tongues.

Though I haven't seen any clear statement to confirm this, I believe this is a very large part of DC's decision to release ZEO. By enabling Zope clusters ZEO addresses pragmatically several limitations inherent in the curent state of Zope / Python. As a side effect, redundancy brings with it answers to concerns regarding reliablility, of hardware, if not Zope itself, in a mission critical setting. While some may see throwing hardware at performance / reliability needs, as inelegant at best, to others with hopes of rolling Zope out for the long term, this seems like a perfectly reasonable solution. Python and Zope will continue to improve, and all Zope users will benefit. For now, leveraging commodity priced hardware is actually about as elegant as it gets. So, DC has heard and answered, and Zope is the better for it!