This week’s newsletter describes a proposal to relay weak blocks to improve compact block performance in a network with multiple divergent mempool policies and announces the addition of five BIP editors. Also included are our regular sections with selected questions and answers from the Bitcoin Stack Exchange, announcements of new releases and release candidates, and summaries of notable changes to popular Bitcoin infrastructure software.


  • Weak blocks proof-of-concept implementation: Greg Sanders posted to Delving Bitcoin about using weak blocks to improve compact block relay, particularly in the presence of divergent policies for transaction relay and mining. A weak block is a block with insufficient proof-of-work (PoW) to become the next block on the blockchain but which otherwise has a valid structure and set of valid transactions. Miners produce weak blocks in proportion to each weak block’s percentage of the required PoW; for example, miners will produce (on average) 9 weak blocks at 10% of the required PoW for every full-PoW block they produce.

    Miners don’t know when they’ll produce a weak block. Each block candidate they attempt has an equal chance of achieving full-PoW, with some of those candidates ending up becoming weak blocks instead. The only way to create a weak block is by doing exactly the same work necessary to produce a full-PoW block, meaning a weak block accurately reflects the transactions a miner was attempting to mine at the time the weak block was created. For example, the only way to include an invalid transaction in a weak block is by risking creating a full-PoW block with the same invalid transaction—an invalid block that full nodes would reject and which would prevent the miner from receiving any block reward. Of course, a miner who didn’t want to advertise what transactions they were attempting to mine could simply refuse to broadcast their weak blocks.

    The high difficulty of creating a 10% weak block and the high cost of creating a weak block with invalid transactions make weak blocks strongly resistant to denial-of-service attacks that might attempt to waste large amounts of node bandwidth, CPU, and memory.

    Because a weak block is just a regular block with slightly insufficient PoW, they’re the same size as regular blocks. When relaying weak blocks was first described over a decade ago, that would mean relaying 10% weak blocks would increase the bandwidth nodes used for block relay by up to 10x. However, many years ago Bitcoin Core began using compact block relay that substitutes the transactions in blocks with short identifiers, allowing the receiving node to only request any transactions it hasn’t previously seen. This typically reduces the bandwidth required to relay a block by over 99%. Sanders notes that this would work just as well for weak blocks.

    For full-PoW blocks, compact block relay not only saves bandwidth but it helps new blocks propagate much faster. The less data (fewer full transactions) you need to send, the faster the remaining data can be sent. Faster propagation of new blocks is important for mining decentralization: a miner who finds a new block can begin working on a successor block immediately, but other miners can only begin working on a successor after they receive the new block through relay. This can give an advantage to large mining pools, creating an unintentional type of selfish mining attack (see Newsletter #244). This problem was common before compact block relay was introduced and it led to both the centralization of mining into large pools and the use of problematic techniques, such as the spy-mining that resulted in the July 2015 chain forks.

    Compact block relay only saves bandwidth and speeds up block propagation when a node receives a new block mainly consisting of transactions it has already seen. But, Sanders notes, some miners today are creating blocks with many transactions that aren’t being relayed between nodes, reducing the benefits of compact relay and putting the network at risk of the problems that existed before compact block relay. He proposed weak block relay as a solution:

    • Miners who create a weak block (for example, a 10% weak block) would casually relay it to nodes. By casually, we mean that the relay would be treated like regular P2P network traffic, such as the relay of new unconfirmed transactions, rather than high-priority traffic like new blocks.

    • Nodes would casually accept the weak blocks and validate them. PoW validation is trivial and would occur immediately; then the weak block would be temporarily stored in memory as its transactions were validated. Any new transactions from the weak block that passed Bitcoin Core’s policy rules would be added to the mempool. Any that didn’t pass would be stored in a special cache, similar to existing caches Bitcoin Core uses for temporarily storing transactions that can’t be added to the mempool (e.g. the orphan transaction cache).

    • Additional weak blocks received later could update the mempool and cache.

    • When a new full-PoW block was received using compact block relay, it could be used with transactions from both the mempool and the weak-block cache—minimizing the need for additional relay time and bandwidth. This could allow a network with many divergent node and miner policies to continue to benefit from compact blocks.

    Additionally, Sanders points to a previous discussion about weak blocks (see Newsletter #173) for how weak blocks could be used to help address pinning attacks and improve feerate estimation. The use of weak block relay was also previously mentioned in a discussion about transaction relay over the Nostr protocol (see Newsletter #253).

    Sanders has written a “basic proof-of-concept with light tests to demonstrate the high-level idea”. Discussion of the idea was ongoing at the time of writing.

  • BIP editors update: after public discussion (see Newsletters #292, #296, and #297), the following contributors have been made BIP editors: Bryan “Kanzure” Bishop, Jon Atack, Mark “Murch” Erhardt, Olaoluwa “Roasbeef” Osuntokun, and Ruben Somsen.

Selected Q&A from Bitcoin Stack Exchange

Bitcoin Stack Exchange is one of the first places Optech contributors look for answers to their questions—or when we have a few spare moments to help curious or confused users. In this monthly feature, we highlight some of the top-voted questions and answers posted since our last update.

Releases and release candidates

New releases and release candidates for popular Bitcoin infrastructure projects. Please consider upgrading to new releases or helping to test release candidates.

  • LND v0.17.5-beta is a maintenance release that makes LND compatible with Bitcoin Core 27.x.

    As reported to the LND developers, older versions of LND depended on an older version of btcd that intended to set its maximum feerate to 10 million sat/kB (equivalent to 0.1 BTC/kB). However, Bitcoin Core accepts feerates in BTC/kvB, so the maximum feerate was actually being set to 10 million BTC/kvB. Bitcoin Core 27.0 included a PR that limited maximum feerates to 1 BTC/kvB in order to prevent certain problems and under the assumption that anyone setting a higher value was likely making a mistake (if they really wanted a higher maximum value, they could simply set the parameter to 0 to disable the check). In this case, LND (via btcd) was indeed making a mistake, but the modification to Bitcoin Core prevented LND from being able to send onchain transactions, which can be dangerous for an LN node that sometimes needs to send time-sensitive transactions. This maintenance release correctly sets the maximum value to 0.1 BTC/kvB, making LND compatible with new versions of Bitcoin Core.

Notable code and documentation changes

Notable recent changes in Bitcoin Core, Core Lightning, Eclair, LDK, LND, libsecp256k1, Hardware Wallet Interface (HWI), Rust Bitcoin, BTCPay Server, BDK, Bitcoin Improvement Proposals (BIPs), Lightning BOLTs, Bitcoin Inquisition, and BINANAs.

  • Bitcoin Core #29850 limits the maximum number of IP addresses accepted from an individual DNS seed to 32 per query. When querying DNS over UDP, the maximum packet size limited the number to 33, but alternative DNS querying over TCP can now return a much larger number of results. New nodes connect to multiple DNS seeds to build a set of IP addresses. They then randomly select some of those IP addresses and connect to them as peers. If the new node gets roughly the same number of IP addresses from each seed it connects to, then it’s unlikely that all of the peers its selects will have come from the same seed node, helping to ensure that it has a diverse perspective on the network and is not vulnerable to eclipse attacks.

    However, if one seed returned a much larger number of IP addresses than any other seed, there would be a significant probability that the new node would randomly select a set of IP address that all came from that seed. If the seed was malicious, that could allow it to isolate the new node from the honest network. Testing showed that all seeds at that time returned 50 or fewer results, even though the maximum allowed was 256. This merged PR reduces the maximum number down to an amount similar to what seed nodes are currently returning.