Question is about participation nodes with at-home networks… For sake of clarity, let’s say a participation node has a billion Algos and proposes a large percentage of the blocks…
Assume all other considerations for bandwidth, storage speed and hardware are acceptable.
If that said home network has an intermittent connection interruption, would the block proposal be delayed or pushed? And if so, is it simply for the half second proposal time?
I guess the question is about resiliency in maintaining < 5 second block time as more and more participation nodes come online on home networks with varying levels of reliability…
Thanks. 
Welcome to Algorand!
This is a very good question.
Actually about 20 blocks proposers are automatically chosen.
Then the voting step of the protocol selects the proposal with the lowest rank.
So there isn’t much issue with downtime for block proposal.
Now, for voting and certifying steps, you need around 80% of the nodes to be honest and working. I don’t know the exact numbers but if you’re too low, you may start selecting committees with not enough good participation nodes and then the proposal gets rejected.
In that case, the block time will increase as there will be a need for another round of block proposals.
This is because Algorand favors security/consistency over availability (the CAP theorem says that you cannot have both consistency and availability in case the network is not perfect). In the blockchain words, we say that Algorand does not fork.
1 Like