The Alpenglow upgrade has been successfully completed! Solana is undergoing a significant restructuring in its history, with a comprehensive transformation of its consensus, economy, and security model.

Author: Frank, PANews

Although it has not received widespread attention, the Solana network has successfully undergone an important consensus and performance upgrade.

On September 1, the Alpenglow proposal (SIMD-0326) of the Solana network was officially passed through community voting. The core achievement of this upgrade is to reduce the network's deterministic block finality time from the original approximately 12.8 seconds to a target range of 100-150 milliseconds. However, Alpenglow is essentially not just a simple parameter adjustment and optimization, but a reshaping of the consensus layer of the Solana network. The implications behind it are far more than just performance improvements; more importantly, it may bring comprehensive changes to the Solana consensus mechanism, economic model, and future development directions. In short, the far-reaching impact of this transformation will radiate throughout the entire ecosystem.

compresses the final confirmation time from 13 seconds to 150 milliseconds, but it's not just about speeding up.

Alpenglow is a new consensus protocol proposal for Solana. It was officially launched by Anza at the Solana Accelerate conference in New York in May. Anza is the team behind the Solana major validator client Agave, as well as several tools and key infrastructure upgrades on the network over the past few years.

The core of Alpenglow lies in the reconstruction of Solana's consensus mechanism, achieving a significant improvement in network performance, and due to the changes in the consensus mechanism, it also radiates and affects the entire network's economic model structure.

Technically speaking, Alpenglow has two core components: the brand new finality engine Votor and the high-performance data propagation layer Rotor.

Before understanding the significant changes brought about by these components, we may need to review Solana's current consensus system, which mainly consists of a system formed by Proof of History (PoH) and Tower BFT. Under the current system, Solana's network confirmation requires two confirmations to achieve a single block confirmation, namely "optimistic confirmation" and "final confirmation."

Among them, "optimistic confirmation" refers to the fact that after a user submits a transaction, they can typically see the transaction status change to "confirmed" within about 500-600 milliseconds. This means that the block containing the transaction has been voted on and recognized by validators with more than 2/3 of the staked weight in the network. However, in reality, "optimistic confirmation" only amounts to a preliminary confirmation, which is theoretically not irreversible. The true, definitive final state, known as "final confirmation," requires a lengthy process. Under the Tower BFT mechanism, a block must reach what is known as the "maximum lock" state, which requires the network to continuously confirm more than 31 subsequent blocks after that block, and the entire process takes about 12.8 to 13 seconds.

In other words, the time for "optimistic confirmation" typically only takes a few hundred milliseconds, but the overall final confirmation time for a block takes about 13 seconds. In such a process, it not only slows down the overall speed of the network but also consumes a large amount of computational resources. Nearly 75% of transactions on the Solana chain are voting transactions.

In the new scheme, Alpenglow's Votor mechanism will completely replace Tower BFT, and the core activities of consensus will be shifted from on-chain to off-chain.

The core change in the Votor mechanism is that validators no longer broadcast on-chain voting transactions. Instead, they exchange voting information directly through a dedicated network. When a block leader collects enough votes, they use efficient BLS aggregation signature technology to aggregate hundreds or thousands of signatures into a tiny "finality certificate," which is then published on-chain as evidence. This process greatly reduces the amount of data that needs to be written into the ledger.

In addition, there is another dual-track voting mechanism within the Votor mechanism. For each proposed block, the network will attempt to reach final confirmation through two paths.

Fast Finality Path (Single Round): If a block quickly receives signatures from validators accounting for 80% or more of the total staked amount, it will be immediately finalized, with a target delay of about 100 milliseconds.

Slow Finality Path (Dual Round): If the signatures collected in the first round of voting are between 60% and 80%, the network will initiate a second round of voting. If the second round also obtains more than 60% of signatures, the block will similarly be finally confirmed, with a target delay of about 150 milliseconds.

In addition to resolving how to confirm blocks and reduce the block ledger, it is also necessary to address the issue of how to quickly send the data required for block confirmation to all validators. Votor is the main mechanism for solving the former, while Rotor is the core component for solving the latter.

The existing Solana uses the block propagation protocol Turbine, which employs a hierarchical tree-like structure to propagate block data, requiring data to pass through multiple layers of nodes before reaching the edge of the network. In contrast, Rotor simplifies this to a single-hop relay model. In this model, the leader splits the block into many small data shards. The leader then sends these data shards directly to a selected group of relay nodes, which subsequently broadcast the data shards to all other validators in the network. This single-hop model significantly reduces the number of network hops required for data propagation, thereby greatly lowering latency.

Consensus mechanism reconstruction, Solana abandons Proof of History (POH)

In this change, Solana will abandon Proof of History (PoH), which was one of the most distinctive innovations of the Solana network.

In the new mechanism of Alpenglow, the efficient propagation of Rotor and the rapid voting of Votor have compressed the block generation and confirmation cycle to within a few hundred milliseconds. At such a short time scale, maintaining a high-precision global clock for continuous cryptographic operations becomes unnecessary and even a performance overhead.

Therefore, Alpenglow adopts a simpler solution: a fixed block time of 400 milliseconds, with each validator independently maintaining a timeout timer locally. If a validator receives the leader's data within the expected time, it votes; if it times out, it skips voting for that slot.

Trade-offs in the changes of economic models and security structures

In addition to the improvements in performance, the new Alpenglow architecture also has a significant impact on various aspects of the economic model.

First, the on-chain voting fee will be canceled. Under the current model, an important cost for validators is the fee for each on-chain vote, which costs about 2 SOL every epoch (2 days). In Alpenglow, a fixed validator admission ticket (Validator Admission Ticket, VAT) is used. According to the proposal, the initial fee is set to be about 1.6 SOL per epoch, which is non-refundable and will be directly destroyed.

On one hand, the design of VAT can reduce the voting transaction costs of validators by 20%, while on the other hand, this destruction can further suppress the inflation of SOL. According to PANews statistics, there are currently about 1,000 validators on the Solana network, so the estimated amount destroyed per epoch is about 1,600 SOL, totaling approximately 296,000 SOL for the year. However, this destruction amount only accounts for about 1.1% of the new supply for that year (calculated based on the current inflation rate of 4.3%).

In addition, reports indicate that after this upgrade, the minimum staking amount required for validators can be reduced from 4850 SOL to 450 SOL. However, this claim seems to lack effective support, as the proposal contents from Alpenglow show that the upgraded Solana network still uses staking to determine the share of validators leading blocks. Furthermore, the specific new staking plan has not yet been announced.

However, in Alpenglow, it is not entirely a case of a fast and secure technological upgrade. Alpenglow reduces the original 33% Byzantine defense cap to 20%, and introduces a "20+20" elastic model, which means that as long as the stake proportion of malicious (Byzantine) nodes in the network does not exceed 20%, the protocol can ensure that no erroneous states (such as double spending) will occur. Based on this, even if another 20% of nodes in the network go offline or become unresponsive due to network issues, hardware failures, etc., the protocol can still continue to produce and confirm new blocks.

Will MEV completely disappear? The 0326 proposal is just the beginning.

In addition to the explicit impact on economic models, the reduction of block confirmation time to 150 milliseconds by Alpenglow also affects multiple ecological roles within the Solana network, with MEV likely being the most significantly impacted.

In the current model, the time window of about 600 milliseconds from when a trade is packed by the leader to when it is optimistically confirmed is the survival space for arbitrageurs or sandwich attackers. Once the confirmation time is drastically compressed, this arbitrage space will be almost completely closed.

Of course, it is not excluded that some MEV participants with top-notch server facilities may continue similar activities, but inevitably, the costs of arbitrage and wrongdoing will also increase significantly.

Additionally, for many existing RPC providers and some projects in the Solana ecosystem, this architectural overhaul may mean that they need to synchronously restructure their own products. Of course, with the improvement in performance, products in the gaming, metaverse, and payment sectors that have extreme performance requirements may have greater room for development.

However, this Alpenglow will be a long process, and the SIMD-0326 proposal that has passed this time is just a very basic plan, merely a proposal that resembles a community confirmation of direction. In the community discussions, it can be seen that there will be a large number of SIMD proposals that will continue to be advanced, such as discussions on whether the specific VAT is confirmed to be 1.6 SOL, or about the relay validator rewards in broadcasting, as well as future staking yield distribution models, and so on.

From the timeline, it is expected that the mainnet deployment of Alpenglow will be completed by the first quarter of 2026. In the community discussions, it is evident that most people are very supportive of this new change. However, some believe that the reduction of voting fees by 20% and the profound impact of MEV may further affect the economic balance of the Solana ecosystem.

Summary

Regardless, with the smooth passage of the SIMD-0326 proposal, Solana's Alpenglow upgrade will continue to advance. It may frequently initiate various key voting activities within the community in the near future. For investors, these votes may impact the future income structure. Throughout this process, there will inevitably be significant engineering challenges and economic games. SIMD-0326 is just the beginning, and whether Alpenglow is the holy grail of performance or a Pandora's box remains to be seen.