Solana is currently advancing a底层 upgrade known as Alpenglow. This version has been enabled in devnet and testnet respectively, which means that both the development team and validators can now check compatibility in the public testing environment, before awaiting the mainnet switch.
The goal is to compress the final confirmation time to 150 milliseconds.
The goal of Alpenglow is to reduce the final confirmation time of transactions from the current system's approximately 12.8 seconds to about 150 milliseconds. By final confirmation, it is meant that once a transaction is completed, in theory, it will not be rolled back.
If this goal is ultimately and stably achieved on the mainnet, the waiting times for exchanges to deposit funds into accounts and for payment applications to confirm receipt of payments from merchants could be significantly reduced.
Two public test networks have been covered.
Anza, which is responsible for the development of the core software Solana, announced on September 25th that Alpenglow had been switched to devnet. The day before that, the independent testnet had completed the transition. The upgrade page for Solana Foundation also indicated that this version was now enabled on both networks.
Among them, devnet is mainly used by the development team for testing software performance with test tokens that have no actual value; testnet is more often used for stress testing network software and verifying the operation of validators.
The adjustment of the consensus method will affect the statistical criteria.
One of the core changes in this upgrade is that verifiers no longer write votes into the blocks; instead, they directly exchange voting information to reach consensus more quickly. According to the design, the network should be able to reach confirmation after one or two rounds of voting.
This change will also have a direct consequence: the total number of transactions in the activity charts for some Solana may decrease. The reason is not that there is a reduction in user payments or transactions, but rather that the votes from validators that were previously counted towards the total number of transactions will no longer be recorded in the blocks. Solana Foundation has informed the data providers to adjust the comparison criteria accordingly.
In addition, service providers that build transaction history records also need to process multiple candidate blocks separately before the final selected block on the network is determined. If the content is mixed in advance, it may result in incorrect records.
The mainnet launch time is not yet determined, but 150 milliseconds remains the target value.
According to the foundation's guidelines, applications that are solely responsible for sending transactions and reading account balances generally do not need to be migrated. However, services that rely more heavily on the underlying data structures and historical records still need to complete adaptation.
Currently, 150 milliseconds is still a target value based on simulation results and has not been continuously verified in a real market environment. Even if the underlying confirmation speed improves, the wallet processing procedures and the exchanges' own fund verification processes may still result in additional waiting times.
Anza has not yet announced the official mainnet launch date. According to their software schedule, feature activation could resume after September 28th at the earliest, but this date has not been explicitly designated as the mainnet launch time for Alpenglow.
