Yesterday the mostly-centralized cryptocurrency Solana suffered a DDOS attack with too many transactions and their validator network got stuck. From my understanding it caused some nodes to miss some transactions and ended up created multiple competing blockchains. You can read about it here. As of right now the network is back up, and the price doesn’t seem too affected, but I suspect this incident will have long term ramifications on Solana achieving widescale trust.
So does Chia share this risk? Sort of, but not really. This kind of issue could cause some annoyances, and maybe even transaction fees on Chia. Maybe even some of the farmers would have issues, but it wouldn’t create the same issue we saw yesterday on Solana. Solana uses really high powered servers as proof of stake validators with a large mempool so they can process transactions extremely quickly. It seemed that some of the validators with the 128GB minimum RAM couldn’t take the high transaction volume from the DDOS that they were missing transactions. It seems like 400 000 were submitted extremely rapidly, and even though Solana can handle 50 000 transactions per second, supposedly, it couldn’t handle that.
From what I understand Chia is not designed to settle transactions as fast as possible like Solana; it is designed to work like Ethereum and Bitcoin and hold transactions until they are settled. If you tried to shove 400 000 transactions into the blockchain it might take you all day to clear through them but it won’t Balkanize the blockchain requiring a restart.
When the Bluebox timelords caused memory issues and a partial DDOS of the network it definitely affected some farmers with low specs or too many connections. But the network was fine. There was only one blockchain. Some people fell behind, and that affected their ability to farm but they didn’t start creating their own version of the blockchain.
I don’t know if this a problem with Solana specifically, and I suspect it might be, or with current proof of stake systems in general, but this class of issue – from problem to coordinated solution – are the result of trying to centralize and accelerate a blockchain to the point where it isn’t the best solution. A high performance non relational database cluster would do far better for what Solana is trying to do. And why Chia’s coin model might not seem fast when compared side to side, but doesn’t create issues like we saw yesterday in Solana. From my understanding as of yesterday if you shove 400,000 transactions into Solana and Chia at the same time, at 50,000 transactions per second to 20, Chia will settle them first. That’s why design matters.