Thanks @bjoernek, @Thyassa has conducted some very detailed research on this group (and others). This isn’t a cluster I’ve got around to adding to my tooling yet though. I’ll do that now, and I’m sure the other guy’s in CO.DELTA will do that too.
Can I suggest that you turn this into a motion proposal?
One thing I’m currently unclear on is, even if the direct incentive to do this is removed, one could argue there is an indirect incentive if we are to make no assumption about the morals of the NPs (which we certainly should not, strictly speaking). Doing this sort of thing increases an NPs chances of gaining control over a subnet.
Shouldn’t the IC Target Topology metrics take this vector into account?
For arguments sake, how can the community inspect and discover that the racks are shared? Or should we assume this is the case if the nodes are in the same DC?