The beauty of memcached is that it isn't on steroids. It just works, and works well from VPS-running-rails-side-project scale all the way up to multi-datacenter-Facebook-platform scale.

You can't push intelligence into your memcached tier because it's all spine, no brain. Exactly how I want my cache to be.

I'm not sure I see what's in here that couldn't be open source within 12 months if enough people actually wanted it. Isn't it a tough sell to say that the open source stacks can't scale, when the majority of the big 2.0 apps are built on them?

Stuff like this is what happens to you when you use .NET instead of Python.

.NET folks have velocity http://msdn.microsoft.com/en-us/data/cc655792.aspx which is actually pretty nice, albeit late in the market

Product placement?