Especially if you live in a pure .NET world, it's a good exercise to see once in a while what's interesting in the Java land. There are some things out there which do not have (not yet, I would add) an equivalent implementation in .NET.
So, if you have around 30 minutes free time, watch out Cameron's interview on theserverside.com.
Distributed, self-replicating caches in the middle tier can be very useful in some enterprise-level scenarios. You meet new challenges especially when (1) you need low latency to access large quantities of read-only data, (2) both your middle-tier and data tier is spread across several layers of machines, and (3) you need a caching solution that scales with the number of nodes.