Stellar Bell: Why the Ledger Reader Needs an Upgrade
Stellar’s Testnet Horizon records a Protocol 29-to-30 boundary on November 3. Mainnet’s vote remains a separate schedule; ledger formats, address routing and signature scope still need different checks.

AI-generated conceptual illustration: a technician fits a reader to a finer clock ring while the coarse scale remains. It represents a data-format change, not actual Stellar equipment or a measured speedup.
Key Points
- SDF’s Testnet Horizon reports Protocol 29 at ledger 5540312 and 30 at 5540313, whose reported close time is November 3 at 17:00:05 UTC. Mainnet’s vote remains scheduled for November 23 at 17:00 UTC.
- SDF expects existing transactions and contracts to remain compatible, but ledger decoders need new XDR definitions. Applications also need to handle the new W address variant where relevant.
- Millisecond timing does not itself establish faster ledgers. ML-DSA verification is for smart-contract logic; network transaction signatures remain Ed25519.
Inside this story
Chronological Timeline
Stellar Core v30.0.0 is published with the Protocol 30 changes. A software release does not establish that a network has adopted them.
The introduction and upgrade guide describe the changes and proposed rollout. The guide was subsequently updated on October 23.
Ledger 5540312 reports Protocol 29 at 17:00:00 UTC; 5540313 reports Protocol 30 at 17:00:05 UTC. These are the service’s ledger close times, not an independently verified vote transcript.
This remains a separate proposed vote. The observed Testnet boundary does not establish Mainnet activation.
An observed Testnet boundary; a separate Mainnet vote
A check after the scheduled vote time on November 3, 2026 found a concrete boundary in SDF’s Testnet Horizon records. Ledger 5540312 reports Protocol 29 and a close time of 17:00:00 UTC. The next ledger, 5540313, reports Protocol 30 and 17:00:05 UTC. The latter’s previous-hash field matches the former’s hash. This is an observed change in the service’s protocol-version records, rather than an inference from the scheduled 17:00 vote.[7][8]
These are ledger close times reported by the SDF-operated public Horizon service, not independently measured wall-clock times for the vote. Our check did not replay consensus or verify individual validator votes. SDF’s introduction and guide still showed their pre-vote text at 18:00 UTC. The guide’s November 23 at 17:00 UTC Mainnet vote remains a separate proposed milestone; these Testnet records do not establish Mainnet activation.[1][2][7][8][9]
Stellar Core v30.0.0 was released on October 16. SDF introduced Bell and its upgrade guide on October 20; the guide shows an October 23 update. Those remain implementation and preparation milestones. In this update, the returned ledger records—not the existence of a release or the passing of a scheduled time—provide the evidence for the observed Testnet protocol boundary.[1][2][3][7][8]
A finer clock can preserve the old meaning of a second
CAP-88 adds a millisecond-resolution close time, closeTimeMs, to the record agreed for a ledger. The familiar closeTime field remains in whole seconds and must equal the new value rounded down to seconds. Reading only the old field therefore keeps the same unit; the proposal does not silently turn an existing seconds value into milliseconds.[6]
Transaction time bounds, minSeqAge, claimable-balance time predicates and Soroban’s ledger_timestamp keep their whole-second semantics. What can change is an assumption around them: consecutive ledgers can, under the new timing rules, share the same whole-second timestamp. “One ledger later” is no longer a sound substitute for “at least one second later.”[6]
The finer clock removes a protocol constraint on possible ledger intervals. It is not a measurement of the network running faster, nor a commitment to a particular new interval. SDF describes Bell as preparation for later steps toward faster ledgers. Clock precision, a chosen operating target and observed performance remain separate facts.[1][6]
The surrounding record needs a new decoder
The boundary is precise. CAP-88 says the ledger that applies the upgrade still uses the older whole-second value format. The next ledger must use one of the new millisecond value types. The format follows the protocol version of the previous ledger, rather than the wall-clock date of a release or an announcement.[6]
XDR is the schema that tells software how to decode these records. The new StellarValue variants also appear inside ledger headers, ledger-close metadata and raw consensus history. A decoder built with older definitions cannot recognize those variants. Even an application interested only in the unchanged whole-second closeTime needs definitions that can first read the surrounding record. This is why the guide requires SDK/XDR updates while describing the transaction-facing changes as additive.[2][6]
A W address carries routing information, not another balance
CAP-84 gives a contract address a 64-bit multiplexing identifier and a textual form beginning with W. The idea extends the existing muxed-account model to contracts: a business can distinguish off-chain customer destinations behind one contract instead of needing a different contract address for each one. The upgraded Stellar Asset Contract transfer function can accept that muxed contract address as its destination.[1][4]
The underlying transfer still goes to the same non-muxed contract. Its event records the identifier as to_muxed_id, so the application can attribute the incoming amount. The identifier does not create an independent on-chain balance or change the transfer’s authorization rules. Customer accounting remains work for the product that interprets this routing information.[4]
There are two reading jobs here too. An updated SDK supplies the W-address parser, but application code must recognize the new variant wherever it validates, displays, stores or routes addresses. Contract argument types matter separately: a plain Address rejects a muxed address, whereas MuxedAddress accepts the muxed form as well as ordinary addresses. A parser update is not proof that every application path has adopted the new destination type.[2][4]
A new verifier does not replace every signature
CAP-87 adds Soroban host functions to verify the ML-DSA-44, ML-DSA-65 and ML-DSA-87 signature variants. A contract can use this capability in its own authorization logic or when checking a signed statement from another system. Existing contracts do not automatically acquire a different authorization policy. The new functions are protocol-gated; importing them before the required protocol version is not supported.[2][5]
Network transaction signatures remain Ed25519. SDF and the proposal explicitly limit this change to verification inside smart contracts. That gives contract-account designers a building block for post-quantum authorization, not evidence that all Stellar accounts or the network as a whole have completed a post-quantum migration.[1][5]
Read compatibility at the layer where it is promised
SDF’s guide says existing contracts keep working, while SDK and raw-XDR consumers need updated definitions and address-handling applications need to consider W destinations. These statements concern different interfaces. They can all be true at once: the accepted behavior of an old transaction can remain familiar while the software that observes it needs a new vocabulary.[2]
For this rollout, three useful questions have separate answers: which protocol the network has adopted, which data format surrounding software can decode, and which optional contract features an application uses. The Testnet records now answer part of the first question. Mainnet adoption and application handling still need their own evidence; a release or a scheduled date cannot answer all three.[1][2][6][7][8]
Continue reading
Explore the topic through other cases and contexts.
Sources & References
- [1]Source 1: Introducing Bell, Protocol 30 on StellarStellar Development Foundation · 2026-10-20Accessed 2026-11-03
- [2]Source 2: Bell, Protocol 30 Upgrade Guide — updated October 23Stellar Development Foundation · 2026-10-20Accessed 2026-11-03
- [3]Source 3: Stellar Core v30.0.0 releasestellar/stellar-core · GitHub · 2026-10-16Accessed 2026-11-03Back to text1
- [4]Source 4: CAP-0084 — Muxed Contract Addressesstellar/stellar-protocol · GitHubAccessed 2026-11-03
- [5]Source 5: CAP-0087 — Host functions for ML-DSA signature verificationstellar/stellar-protocol · GitHubAccessed 2026-11-03
- [6]Source 6: CAP-0088 — Millisecond-Resolution Close Timesstellar/stellar-protocol · GitHubAccessed 2026-11-03
- [7]Source 7: SDF Testnet Horizon — ledger 5540312, reported Protocol 29Stellar Development Foundation · Testnet HorizonAccessed 2026-11-03
- [8]Source 8: SDF Testnet Horizon — ledger 5540313, reported Protocol 30Stellar Development Foundation · Testnet HorizonAccessed 2026-11-03
- [9]Source 9: Stellar Networks — SDF Testnet and its public Horizon serviceStellar Development FoundationAccessed 2026-11-03Back to text1


