Are you in support of blacklisting IPs as in XT?
- Thread starter Smoothie
- Start date
IMO it could be done better but, a) I wouldn't really call it a blacklist, rather a priority scheme and b) Mike's right in his statement that TOR unfortunately is used for a lot of attacks.
I think Mike explained it well here: https://www.reddit.com/r/bitcoinxt/comments/3in4qo/mike_hearn_shares_his_initial_thoughts_on_bip100/
I think Mike explained it well here: https://www.reddit.com/r/bitcoinxt/comments/3in4qo/mike_hearn_shares_his_initial_thoughts_on_bip100/
What you're blocking is not IP's of fellow bitcoin peers, but IP's of attacking bot nets coming through publicly listed Tor nodes. And that's only if your connections get maxed out from an attack. Then, your node just prioritizes real bitcoin peers. Sounds good to me.
I was initially concerned (I read the 'headline') but when I studied the actual body of work it seemed fair enough. So I refrained from switching to the 'big blocks only' tag, as it seemed to me to be a reasonable first stab at something that is a potentially desirable feature. The reality is pure freedom is just as unsustainable as pure censorship. We have to find a balance imho.
So in that regard I'm alway worried about the creeping nature of change, as cracking the door can sometimes lead to opening the floodgates - but I'm not so scared as to automatically reject anything like this for fear of what it might become. Things have to be judged on individual merit. This particular 'feature' seems to have merit.
The best rationalisation I thought was, that if it was to get hairy then what is to stop people forking XT. Evidently that is a philosophy that XT devs back and so whats good for the goose is good for the gander.
So in that regard I'm alway worried about the creeping nature of change, as cracking the door can sometimes lead to opening the floodgates - but I'm not so scared as to automatically reject anything like this for fear of what it might become. Things have to be judged on individual merit. This particular 'feature' seems to have merit.
The best rationalisation I thought was, that if it was to get hairy then what is to stop people forking XT. Evidently that is a philosophy that XT devs back and so whats good for the goose is good for the gander.
I read that it only kicks in when overloaded. In that case seems ok to me, although some non-centralized determinant would be better.
Well, a fancy self-learning scheme (machine learning maybe even to figure out rogue peers?) would certainly look nicer - I think it is in the category of 'good enough, given limited time to work on all things.'
It should have not blown up into this huge issue, but I guess that is simply reflecting that people doing real work are directly interacting, and not some big corp. PR departments. It creates (price) volatility and some heat, but otherwise also shows that the environment is still young and full with people who actually give a damn.
It should have not blown up into this huge issue, but I guess that is simply reflecting that people doing real work are directly interacting, and not some big corp. PR departments. It creates (price) volatility and some heat, but otherwise also shows that the environment is still young and full with people who actually give a damn.