Do we know which BCH implementations did the best and which did the worst during the stress test? Tom Zander said Bitcoin ABC did the worst. How is performance gauged?
Only metric there is, is the fallout of 250 BC nodes at the time of the 'test' - https://cash.coin.dance/nodes/all. Maybe they were all low-end systems. I wouldn't make claims like Tom
That >25% of ABC nodes got knocked down is highly concerning. No statement from ABC or miners as to why. BU nodes all held up. Maybe we do need to run more stress tests after all.
In that case, we need to test again before 11/15. If users could result in fallout of enough nodes, would we be “voting” for an implementation indirectly without hash?
I've been saying since 9/1 that the next stress test should be 10/1 and so forth. There were too many kinks to work out and real volume can come in very very quickly.
I'm not necessarily saying sustained volume at those levels but in a bull market you can get media hits that sebe usage into a frenzy. We don't want to screw the pooch at that time.
From 25,000 to 2.5M daily tx are two orders of magnitude. 100x. BCH as is can take any media hysteria in the short term (2 years). And easy fix of burst relay gets us to 8M daily tx.
Aright I'm with you there. Although there is the so called "fidelity effect" where you want to be able to show bigger buisinesses that this thing isn't going to choke under pressure
and just removing the limits is just a PR, or even outright lie, if one wants to use that as a proof that it won't choke. Software needs optimization, and removing the bottlenecks.
I don't think that's what anyone is claiming. The "remove of the limit" rhetoric acknowledges that miners wont try to propagate a block that has a high probability of getting orphaned.
But is it. I saw many times 'fidelity problem' and rhetoric 'remove the limit to show the chain can scale'. Remove the bottlenecks, that's the way to show haw the chain can scale.
Simply, software is not optimized and can not use hardware resources, even if limits are removed. There are many bottlenecks to be removed before we can show BCH won't choke.
And BCH is not yet there - with current software implementations it will choke under >8M daily TX volume, no matter if the block limit is even removed, or not.
I've been saying since 9/1 that the next stress test should be 10/1 and so forth. There were too many kinks to work out and real volume can come in very very quickly.
It could be small rasp pi nodes which got knocked out. It would only be an issue if it hit the miners. Also, you can say that 75% of the nodes passed the test - Glass half full/empty?