I like DSLs. Here’s a paper about one, “Experience with a Language for Writing Coherence Protocols”
It’s slightly amazing to me that coherence protocols are a rich and complex enough subject to merit developing a DSL. Reading the article points out why such is true, though.