Manik, can you explain the difference between NodeSPI and WorkspaceNode? Why are they in parallel? Another question. if I stick with using the operation from Cache interface only, will there be any implication for performance reason for the new proposed change? Finally, from this hierarchy, I still don't see a clean way to access CacheSPI (or NodeSPI), except to do: CacheSPI spi = (CacheSPI)cache; But this may subject to break in the future. Should we simply add a factory method for the SPI instance? View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3994641#... Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&a...