Hi!
Most certainly. The whole idea of TPF is based on caching. We earlier had a discussion on “Web-grade”, and designing for cache efficiency is an important part of that. Also see work on decentralized caching in this area (http://linkeddatafragments.org/publications/#caching).
Thanks, the CyCLaDEs paper appears to be paywalled, but I certainly will read the other ones.
I don't think so, if caching is enabled. E.g., we can just put a CloudFlare cache in front of this, and it will work*.
I'm not sure how effective "naive" cache like CloudFlare would be when we're talking about billions of triples in various combinations. TPF requests are pretty elementary AFAIK and wouldn't that mean that CF would have to pretty much load every view on the graph engine index and store it in every format (assuming content negotiation issue is solved)?