It occurred to me that just saying something is possible is pretty far from demonstrating it. So I’m going to pretend this is a software blog for just a moment. If you haven’t been following my last couple of blog posts, this isn’t the place to start.
Somebody mentioned, out on the intertoobz, that my distributed-content idea is nice, but it’s easier to find content in a centralized system. To which I say: distributed content can have a centralized index. So, the drop-dead simple architecture is a server farm that does nothing but redirect data requests to individual machines, scattered wherever. Users wouldn’t see much difference between that and a standard website. In fact most users wouldn’t see any difference at all.
To make it more fun, the index can be widely distributed as well, with many “central” nodes. I stole an idea from Mike Nolet at AppNexus for another project, and essentially it does load-balancing via a hacked DNS server. It rotates IP addresses through. A version of which is sitting on my laptop right now–it just does a round-robin thing, but that logic could be improved upon to weight its IPs according to capacity/performance/uptime/whatever.
There are a couple of approaches to ensuring the distributed content servers are online & functional. One is regular polling (which my Scarecrow application already does in another context, so that’s easy) and the other is for a machine to request work when it’s free…which could be implemented via RabbitMQ or something similar. If it’s RabbitMQ, again my Scarecrow app already does it, because that’s how it distributes work to virtual servers in different physical locations.
I guess I’m saying I already have many pieces of a functional system architecture lying around as leftovers from something or other. Which doesn’t at all mean I could get (literal) buy-in from authors, and it sure doesn’t mean I’m the right guy for marketing this or anything else. Mostly, when I talk people look for reasons to disagree, so it’s better if I don’t talk.
Anyway. It’s a neat idea, and I think it’ll happen sooner or later whether I exert myself further or not. Lends itself well to access via mobile apps, too.
The thing is, the current ebook distribution system is horribly inefficient from an author’s point of view. If it had started in the software world, it probably wouldn’t at all resemble what people currently think of as normal. So there’s lots of room for improvement, and the limiting factor is probably nothing more (complex or simple, pick one) than an approach’s credibility with indie authors.
If two of you have gotten to this point, I’m going to throw myself a party. Maybe putting the idea out there like this will help to get it out of my brain?
Back to our irregular programming next time. If I gave you a headache, I apologize. Try lying down.