> Registries generally don't need the durability that comes with RDBMS, > and unless you are already outfitted with clustered RDBMS tech and the > resources to management it, the complexity is likely not worth the > trouble.
I've really warmed up to that view. The use of "Remember Me" might be a counter argument, but as I said we don't use it and don't plan to in the foreseeable future. > given the proliferation of hardware virtualization and the > redundancy and vertical scaling they often provide, my default > recommendation is a single node CAS deployment with in-memory Ticket > Registry and active/passive configuration for disaster recovery. The point about virtualization pretty much invalidates the argument about active-active setups making more efficient use of resources. (The consideration of the cost of system administration/patching on hosts that are unused 99.9% of the time may yet be fair.) > For situations that insist on a multi-node CAS deployment, my default > would be to go with a distributed in-memory Ticket Registry like > Ehcahce. The use of Terracotta underneath a distributed Ehcache instance is the showstopper for me. It smelled of magic and mystery in my brief experience with it, which are smells I associate with headaches in production. I'm heavily leaning toward memcached. (Sans repcache if it matters.) M -- You are currently subscribed to cas-dev@lists.jasig.org as: arch...@mail-archive.com To unsubscribe, change settings or access archives, see http://www.ja-sig.org/wiki/display/JSG/cas-dev
