Zilliqa, Ethereum и Radix
Относно Zilliqa Breaks 2,400 Transactions per Second with their ‘Sharding’ Blockchain
http://www.the-blockchain.com/2017/10/1 ... lockchain/
Zilliqa - 3600 възли постигат 2440 tx/sec
Ethereum - 24000 ( https://www.ethernodes.org/network/1) възли и ако ползват същата технология биха достигнали 15000 tx/sec
Radix - при 22000 възли и ниво на излишък 100 се очаква скорост от 220000 tx/sec
Колкото по-малко е числото на излишъка, толкова по-голяма е вероятността възлите обслужващи раздела
да бъдат изключени и да престане обслужването на раздела. Вероятността 100 възела обслужващи раздел
да се изключат е минимална. За теста с много раздели ще се приеме, че даден раздел се обслужва от 10 възли.
feserleer:
say we have 22k nodes, and we want a redundancy ratio of 100
thats 22000 / 100 = 220 * 1000 (assuming 1000 tps capability per node) = 220,000 tps
that 100 redundancy ratio really turns into almost 200 (as a transaction will be stored in at least 2 shards the majority of the time)
well, the lower the redundancy, the more chance of all nodes serving a shard go offline and the data of that shard is very difficult to recover, or not at all
likelihood of all 100 nodes serving a shard going offline before the network notices a shard is under served is pretty damn low
(ignoring malicious actors ofc...thats for another day)
we could probably get away with less, but we should aim for as high a ratio as possible
for our testnets I'm going to tune it to about 10...so we can extract the most out of it