HALLMARK WEIGHTING PROBLEM
ZjP and I were discussing the implications of the hallmark weighting algorithm on blockchain propagation. We believe that the hallmark weighting algo is actually detrimental to the network staying synced and preventing DDoS attacks. The nodes that are marked with higher balances are being repeatedly bombarded with requests for the blockchain even if they are already saturated with requests and non-responsive. The smaller balance or zero balance nodes never try to sync with nodes which have a lower hallmark weighting or no hallmark weighting. This results in the higher weighted nodes, which are currently the backbone of the system, being overrun and taken out. If you are running non-hallmarked servers to try to boost the network's stability, you are actually causing more damage because you are straining the already overloaded nodes that have a large weight.
If the hallmark weight system were to be removed, the network would view each node, regardless of it's balance, as equal and distribute the workload across entire network, instead of concentrating the entire network's requests on a few heavily weighted nodes. This would mitigate the DDoS attack because they be forced to try to take out more targets instead of being able to only target a few large balance nodes and bring the network to a halt.
Also, it seems that whenever the blockchain gets a bad block, you have to delete and re-download the entire blockchain to resync. A proper resync only occurs if the blockchain pauses, but doesn't download a bad block.
Come-from-beyond, can you please relay this information to BCNext? What are your thoughts? I hope the source code is release asap, so these problems can get sorted out quickly.
ZjP and I were discussing the implications of the hallmark weighting algorithm on blockchain propagation. We believe that the hallmark weighting algo is actually detrimental to the network staying synced and preventing DDoS attacks. The nodes that are marked with higher balances are being repeatedly bombarded with requests for the blockchain even if they are already saturated with requests and non-responsive. The smaller balance or zero balance nodes never try to sync with nodes which have a lower hallmark weighting or no hallmark weighting. This results in the higher weighted nodes, which are currently the backbone of the system, being overrun and taken out. If you are running non-hallmarked servers to try to boost the network's stability, you are actually causing more damage because you are straining the already overloaded nodes that have a large weight.
If the hallmark weight system were to be removed, the network would view each node, regardless of it's balance, as equal and distribute the workload across entire network, instead of concentrating the entire network's requests on a few heavily weighted nodes. This would mitigate the DDoS attack because they be forced to try to take out more targets instead of being able to only target a few large balance nodes and bring the network to a halt.
Also, it seems that whenever the blockchain gets a bad block, you have to delete and re-download the entire blockchain to resync. A proper resync only occurs if the blockchain pauses, but doesn't download a bad block.
Come-from-beyond, can you please relay this information to BCNext? What are your thoughts? I hope the source code is release asap, so these problems can get sorted out quickly.
hmmm probabilly into the actual situation, removing of the hallmark system is the best solution... isn't?
