Hassan Maishera
TL;DR
-
Solana’s Alpenglow consensus upgrade is moving to the network’s public testnet for broader testing.
-
The upgrade aims to reduce transaction finality from approximately 13 seconds to 150 milliseconds.
-
Alpenglow replaces TowerBFT with Votor, allowing validators to finalize blocks after one or two voting rounds.
-
Users and applications will not need to change their wallets or transaction procedures.
Solana developers are moving the Alpenglow upgrade to the blockchain’s public testnet as they work toward reducing the time required to finalize transactions from approximately 13 seconds to just 150 milliseconds.
The testnet deployment marks a significant step for an overhaul that could make Solana payments and applications respond substantially faster. It follows more than four months of testing on a smaller network created specifically for Alpenglow.
Moving to the established testnet will allow developers to determine whether Solana’s wider collection of validators, services, and infrastructure can migrate to the new consensus system successfully before any potential mainnet activation.
Alpenglow Targets Near-Instant Transaction Finality
Finality is the point at which a blockchain transaction becomes effectively irreversible. Reaching it confirms that a payment or other operation is part of the network’s permanent record and will not be removed by a competing chain history.
The concept has practical implications for many services built around a blockchain. Cryptocurrency exchanges often wait for finality before crediting customer deposits, while bridges need confirmation before releasing corresponding assets on another network. Merchants also require assurance that an accepted payment cannot be reversed.
Solana currently achieves finality in approximately 13 seconds. Alpenglow is designed to cut that delay to about 0.15 seconds, or 150 milliseconds.
If achieved under real-world conditions, the improvement would make finalized blockchain transactions feel nearly instantaneous. It could benefit payments, decentralized trading, and other applications where even short delays affect the user experience.
The target represents an intended performance level rather than a guarantee. Testnet trials will help determine whether the protocol can maintain that speed while operating across a broad and geographically distributed validator set.
Votor Replaces Solana’s TowerBFT System
Solana currently uses a consensus mechanism called TowerBFT to help validators agree on which transactions should form part of the permanent blockchain record.
Under that system, validators record their votes onchain and accumulate enough confirmations across 32 slots before a block reaches finality. The process provides security but creates a delay between the initial execution of a transaction and the point at which it becomes irreversible.
Alpenglow replaces TowerBFT with a new voting protocol called Votor. Instead of placing the entire sequence of validator votes on the blockchain, Votor allows validators to communicate those votes directly with one another.
The network can then settle on a block after one or two voting rounds. Removing the long chain of onchain votes is expected to produce Alpenglow’s substantial reduction in finality time.
The upgrade changes Solana’s underlying consensus process without requiring application developers or ordinary users to adopt a different transaction system. Applications will continue executing transactions in much the same way, while users will not need to replace their wallets or alter how they send funds.
Solana’s public testnet functions as a copy of the blockchain whose tokens do not carry monetary value. It allows developers and validators to test software upgrades without putting users’ assets at risk.
Participants can rehearse the Alpenglow migration, observe its behavior under more realistic network conditions, and identify problems that were not visible during smaller-scale testing.
Developers can also restart the testnet if a serious fault occurs—an option that would be highly disruptive and potentially costly on the main network.
Alpenglow has already spent more than four months operating on a purpose-built testing network. The public testnet phase raises the stakes by introducing the protocol to more of the computers, software, and supporting services that already participate in the Solana ecosystem.
A successful test will require those different components to switch to the new consensus process in a coordinated manner.
Test Requires Agave 4.3 Validator Software
Validators participating in the testnet migration must run Agave 4.3, the latest version of Solana’s primary validator software.
Agave is maintained by Anza, one of the development companies contributing to the Solana network. Anza recommended the new version to all mainnet validators on September 21 after conducting a phased rollout.
The software was initially introduced to operators responsible for approximately 10% of the SOL used to secure the network. That share was subsequently increased to 25% before Anza issued its broader recommendation.
A gradual deployment allows developers to observe the software’s performance among a limited section of validators before encouraging adoption throughout the network.
However, two alternative Solana validator clients—Firedancer and Frankendancer—do not yet support the Alpenglow test. Both programs are being developed by Jump Crypto as alternatives to Agave.
Multiple validator clients are important because they reduce a blockchain’s dependence on a single implementation. If all validators run the same software, one critical defect could potentially affect the entire network.
Because Firedancer and Frankendancer are absent, the first Alpenglow testnet migration will depend entirely on Agave. Support from alternative clients may therefore become an important consideration before a wider mainnet deployment.
Anza’s schedule lists September 28 as a tentative date for activating features included with Agave 4.3 on Solana’s mainnet.
That entry should not be interpreted as a confirmed date for launching Alpenglow on the live blockchain. The broader upgrade remains in testing, and Anza’s tracker still listed the testnet activation as pending early Wednesday.
Developers will first need to assess the public testnet migration and resolve any technical problems it uncovers. A mainnet launch will depend on the results of that work, validator readiness, and coordination across the broader Solana ecosystem.
This latest development comes a week after Solana tripled its maximum transaction size to 4,096 bytes, giving developers more capacity for complex trades, multisig wallets, and privacy apps.
