But... But I was told non-mining nodes are essential to Bitcoin security and decentralizationings 
And I when I told you that non-mining nodes are irrelevant, because trivially faked, you got upset...
And now ...you're telling me that non-mining nodes are irrelevant, because trivially faked?
And I when I told you that non-mining nodes are irrelevant, because trivially faked, you got upset...
And now ...you're telling me that non-mining nodes are irrelevant, because trivially faked?
Non-mining nodes are essential to security and decentralization. That doesn't mean that we have good (or complete) information about node counts. The former is a statement about how miners' selfish interests are balanced.
All the
That's a lot harder to argue. I said data on nodes is incomplete -- not altogether useless. Operating a node means expending resources on bandwidth, storage -- those have costs. That's why the key to this data is in short term vs. long term analysis. It would be astronomically more expensive to operate 5000-8000 Core nodes over the past two years than it would be to spin up 800 Classic nodes over the past week. That the data is not completely transparent does not mean we should ignore it entirely.
The latter is just an acknowledgment that nodes can be Sybils -- that node count, particular over the short term, is not the most realistic measure of node proportions among incompatible softwares. One indication that a Sybil attack is occurring right now: comparing to overall node counts 1 month ago, when there were no Classic nodes, 65% of Classic nodes are "new", i.e. not Core nodes switching to Classic, which is contrary to the larger trend in node health. The picture is further obscured by that fact that pseudonode (NotXT) was released for Classic, so presumably people are also actively spoofing Classic nodes with the intention of later shutting them down.
TL;DR: Bitcoin relies on "nodes [...] essential to security and decentralization," a mechanism both trivial to bypass/defeat by creating legit nodes on VM instances & spoofing.
Faith restored, sinking all my money into this shit
Faith restored, sinking all my money into this shit
Nothing here indicates that the node consensus mechanism is trivial or easy to bypass. Pseudonodes aren't even operating with the new fork's consensus rules. What we're saying is that the numbers can give the appearance that Classic has significant node support even if it doesn't.
This is why node centralization is relevant. What used to be over 10,000 nodes is now closer to 5,000. The more bandwidth pressure we put on nodes, the less nodes we have, and the easier it is to mount a Sybil attack, such as the one Classic backers are currently mounting.