The Solana outage has once again raised questions about the design of major layer 1 blockchains.
Validators attempted to restart the network twice over the weekend after on-chain activity froze. For a while, Solana was only doing 93 TPS. After performance degradation, the validator community eventually had to elect to restart the network's sync chain. According to the official blog post, the cause of the outage has not yet been determined, but Solana is still actively investigating. The team stated that without the intervention of the validator community, the network would not be able to recover.
Engineers debugging the problem suggested restarting the network. Following this, the validator community collectively decided to downgrade to the previous stable version, v1.13.6. At the same time, 1.14 was rolled out to reduce restart risk. Last year, Solana experienced 14 outages, ranging in length from time to time. Angry community members pointed out that the ongoing outage could be due to a huge design flaw in the network. A user named “DBCrypt0” stated that Solana’s on-chain consensus model involves network transactions consisting of consensus communication between validators and the transactions themselves, which “increases transaction volume and TPS.”
They further criticized the network, saying that out of the 4k TPS owned by Solana, only 10% were "actual" transactions, while adding that most of the transaction volume was made up of messages from validators, which ultimately bogged down the system.
In the event of an outage, communication between validators stops, during which time they turn to Discord to make plans. However, two-thirds of the validators must agree on a solution for recovery, some of whom may be offline or unaware of the outage.
Solana's technical slowdown is nothing new. Last year, its development team even worked out ways to mitigate such incidents. Recently, developers implemented a new feature called "Priority Tolling," which enables users to pay extra to avoid congestion.
