dcposch.eth
7mo
Real engineering culture difference. Here's the founder of Solana tweeting 150+ hours of mainnet downtime, including 4.5 hours last year--cause undetermined (!)--as if NBD. "We'll get to three 9s of uptime in a decade."
Ethereum has zero downtime since launch.
The closest Ethereum's had to downtime is some large gas spikes, but that's:
1. unavoidable in a permissionless protocol
2. still way better than even 10m of hard downtime, let alone 4h+
3. mitigated by upcoming L1 scaling
2
3
30
No contradiction. L1 scaling makes gas spikes less severe and less common, but in a permissionless protocol it is always ultimately an open auction.
If you have millions of dollars' worth of priority fee in the mempool fighting to get onchain, you either pay or wait.
1
1
It cannot be completely solved. Every chain has *some* throughput limit less than infinity. Then if it's public & permissionless, what do you do if one day there's a huge backlog of senders willing to pay $100 a tx?
You either wait or pay $101. No option C.
Every major Solana outage is followed by a “fix” that just makes the network more centralized
SWQoS is a great example after the 2022 downtime, they solved it by giving more network access to big stakers
1
2
Nah, 6 nines is exotic high availability. It means ~30 seconds of downtime a year, so if your API is down for 5 minutes once you've blown your "6 nines" downtime budget for the decade.
4 is more standard.
1