TGViewer
TON Core TON Core @toncore · 6.88K subscribers
Post #98 11.9K
The R&D phase has been completed — research and development of the task of speed up operations in the TON blockchain to less than 1 second while maintaining scalability, security, and decentralization.

During the work, three key areas were identified that, taken together, will allow the desired result to be achieved.

1. L1 and Node improvements

1.1. New Catchain 2.0 consensus

The current Catchain + BCP consensus stack has been replaced with a new protocol — Catchain 2.0, based on Simplex / Alpenglow algorithms adapted to the TON architecture.

The consensus protocol determines how validators share data, verify and accept blocks with transactions.

Simplex and Alpenglow are modern open consensus protocols that are characterized by high speed and relative simplicity.

Catchain 2.0 allows blocks to be finalized in 200–400 milliseconds while maintaining the same level of security. Previously, the block finalization time was about 2.5 seconds.

An additional advantage in terms of security is the widespread knowledge of these protocols, which greatly simplifies their audit.

1.2. New block broadcast protocol — "two-step" broadcast

Blocks are transferred between network nodes using the broadcast protocol. We have redesigned this network protocol, reducing the block delivery delay from approximately 700 ms to ~100 ms.

1.3. Additional optimizations

A number of additional optimizations have also been made to the C++ implementation of the TON node, including:

— block compression during transmission;

— parallelization of validation;

— the ability to exchange data between validators not only via the RLDP2 protocol, but also via TCP or QUIC. Based on the test results, it may be possible to switch to one of these protocols.

In addition to the main solutions, the internal statistics system was refined, as well as the internal tools for benchmarks and measurements for nodes, validators, and API.

2. API

2.1. Reduced API latency

The Toncenter API has been optimized, resulting in a latency of 30–100 milliseconds for obtaining the status of an operation via the Streaming API V2.

2.2. New functionality

Streaming API V2 in Toncenter now returns two statuses of operation instead of one:

confirmed — the block has appeared in the shardchain;

finalized — the block has been accepted into the masterchain.

Previously, only the finalized status was supported.

3. UX

A new UX approach has been proposed. Since an operation with the confirmed status has a rollback probability of less than 1%, applications can display it to the user at this stage. The user can then immediately perform the following operations without waiting for finalization.

At the finalized stage, the application can additionally mark in the UI that the operation is completely finalized.

This is similar to the confirmation model in other blockchains, where the operation is displayed after being included in the first block, and final finalization occurs after several block confirmations — at the same time, the user can immediately continue interacting with the blockchain.

This is a solid addition to last year's UX and kernel updates.

All of the above components have been completed separately. Component integration is underway, followed by testing on the TON test network.
  • 🔥 63
  • ❤ 39
  • 👍 34
  • ⚡ 11
  • 🦄 6
  • 😘 2
  • ❤‍🔥 1
  • 🤩 1
  • 🤨 1
More from @toncore
  1. Sep 14, 2026Updated staking smart contract lineup Single-nominator (v1.1) — Direct validation with one…
  2. Sep 14, 2026Nominator Pool 2.0 Nominator pools let Gram holders stake funds to support network validat…
  3. Sep 7, 2026LLM-based content moderation in Toncenter API Toncenter API now automatically filters inap…
  4. Sep 5, 2026Pending API and Streaming API Performance Improvements We’ve shipped a major performance u…
  5. Sep 5, 2026TON performance update With the August 17 update, TON mainnet validators transitioned to a…
  6. Sep 5, 2026Service updates minter.ton.org — Updated to jetton-2.1-tolk, the latest jetton smart contr…
Threads Profile ViewerView any public Threads profile without an account.Open ThreadLook →Writing with AI? Make it sound human.Metric37 rewrites AI drafts so they read naturally. Free AI detector, 1,500 words free.Try Metric37 →