So having learned more about this -- doesn't this rely on relays pretty significantly to host book parts "in perpetuity"? I mean I guess this is a broader question for #nostr in general, that if the network achieves any kind of scale, relay storage needs are going to blow up?
Discussion
Not really, since events are atomic and can be downloaded and stored on any computerized device running a relay-websocket or relay-API (RESTful).
It's widely-distributed storage, but the storage doesn't need to be very redundant. There are already thousands of relays, all over the globe. You can spread the publications out. You don't need to have each publication on every relay, you just need each one on several relays, as clients can crawl Nostr and find them, wherever you put them. Even if that's just your laptop running a webserver.
Nostr actually solves the Big Data storage problem.