I saw your question there but I have no answer for it. Yet. :wink: Here are some thoughts though...
We've been discussing p2pool-like solutions ("a secondary, short blockchain", "it holds a month's worth, or 3 rounds worth and dumps the older info to keep it short and sweet") inside dev team some time ago but hadn't yet come to conclusion if it really suits us afaik. Problem with another blockchain - someone has to mine it. And hashrate there must be enough to be at least as secure as main chain. We can make it merge-mined with main or smth like that however but it means same miners will produce consensus for sharing their own block reward with masternodes. And there's something wrong with that model imo. Or we can make it POS-like but I'm not sure how exactly. Also Evan proposed another model some time ago - storing limited MN list changes in blockchain or smth like this if I remember correctly. The idea there was that we can make every node on the network to be aware of the exact same MN list at any given point of time by utilizing the same blockchain. I'm still not sure which approach is better however.
I see no significant gain from "several masternodes functioning per round" other than making payments more even so people could sleep better :grin: for example for 4 MNs per block you'd get ~ 0.5 DASH almost every day at current MN count. Other than that.... I don't know. As for blockchain bloat - in that example with 4 MNs it means 5 outputs instead of 2 for "normal"-mined block and say 103 instead of 100 for p2pool-mined block (take this one for example
http://explorer.dashninja.pl/tx/76ef1856f2c777c26adf95b90db9077bb4986d4079f150f38d67e510c4092f23 - 141(!) outputs). Not too much bloat imo.