I'm confused how you find the data again in this cache? It seems like you're randomly salting your cache keys, so I don't understand how subsequent requests find cached data.
Wish this was clarified. One thought would be to choose a stable component to in turn hash, such as the source coordinate hexagon, but then why not just put that in the hashtag
Naively (and naive is often fine), you can query each node in parallel. The number of nodes is usually very small relative to the amount of data per node and reading is usually much faster than writing.
I see them pretty frequently.
How'd I guess on the article itself?