If it is the hosting cost that is the factor here then aren't there quite good distributed options (Bittorrent, Dat, IPFS) out here nowadays that would simply l

If it is the hosting cost that is the factor here then aren't there quite good distributed options (Bittorrent, Dat, IPFS) out here nowadays that would simply leave hosting up to who ever wanted to store the paper. Then all you need is a frontend for the hosting. Arxiv or whatever could serve that nicely. After that you can include these in whatever journal you want. Sorry maybe I am missing something. I've never really felt like I understand scientific publishing, neither the traditional journal model or the Arxiv model.