NOTE: my concern here is to do with network health and stability and trying to find a helpful solution to any large node operator’s perceived problem.
The problem being solved is that the people running mega machines, and mega farms do not like to upgrade nodes. Typically some will just turn off (kill directories or whatever) and start their nodes afresh with the new version. For them its easy and relative fast. Well faster and earn more than other methods
Doing a cold turkey upgrade is bad for the network and people’s mental well being (those who follow/need the emissions).
Some have suggested a 2 or 3 day grace period which is good in theory, but if a node operator is reluctant to upgrade then they may very well wait till the last minute or maybe last 12 hours.
My idea is to use the fact there is 4 emissions machines & there being a 12 hour cycle which allows a staged path to switching over which version is paid and encourages people to upgrade quickly but gives them time to do it as well. The all or nothing approach has proven to be a nightmare for the network.
2 day grace period (4 days would work easy enough too)
- 1st step is to tell one emissions machine to only give emissions for the latest version, and leave the others as is
- 2nd step is tell the next emissions machine to give emissions for only the latest version
- 3rd step is to do the 3rd machine
- 4th is the 4th machine
These 4 steps could be done every 12 hours or be done one a day. And is a very simple easy way for Autonomi to implement this. No coding needed really since they already tell each emissions machine which version(s) to give emissions for.
This way the mega operators will know they still will receive emissions for the nodes yet to be upgraded. For some of them it takes 24 hours for one of their machines to upgrade using the correct method of restarting (upgrading) one node at a time while leaving the others running. And if they know the older nodes will still receive some emissions then they are incentivised to do their upgrading the desired way rather then stopping all at once and restarting from scratch with new directories etc.