Published on: 2015-07-24T03:54:31+00:00
In a 2015 discussion on the bitcoin-dev forum, a member raised concerns about the potential discrepancies and errors that could arise from downloading a random selection of blocks from each peer. This highlighted the importance of ensuring consistency and accuracy in blockchain data for the functioning of decentralized digital currencies like Bitcoin.On the bitcoin-dev mailing list, there was a conversation about improving the speed of bitcoin node connections for syncing purposes. Matt Corallo suggested running servers around the world and testing based on closest-by-geoip to improve throughput. However, he also pointed out the potential issues with sybil resistance. Further experimentation was planned to explore possible improvements.In an email exchange, Marcel Jamin and Peter Todd discussed the idea of measuring the upload capacity of peers by downloading from them. The context or background for their discussion was not provided.Another post on the bitcoin-dev mailing list focused on testing the Bitcoin network's maximum throughput to determine if it could support a faster block rate. Nodes were randomly selected from peers.dat, and a random selection of blocks was downloaded from each peer. The results showed that 37% of connected nodes failed to upload blocks faster than 1MB/s, indicating that the network did not have sufficient bandwidth for increased block sizes. The findings also revealed that 16% of nodes uploaded blocks faster than 10MB/s. These results emphasized the challenges in increasing block sizes and the need for careful evaluation.The same topic of the network's available bandwidth for increased block sizes was discussed in another email thread. One participant mentioned Gavin's numbers, stating that 37% of nodes would fail to upload a 20MB block to a single peer in under 20 seconds. Another participant pointed out that the measurements were optimistic as they only measured download capacity, which is further reduced when sending blocks to multiple peers for reliability. It was noted that additional margin was needed to resist attack as performance is consensus-critical.A test run on the Bitcoin network was conducted to assess its ability to support a faster block rate. Randomly selected nodes were contacted, and a random selection of blocks was downloaded from each peer. The results showed that 37% of connected nodes failed to upload blocks faster than 1MB/s, while only 16% uploaded blocks faster than 10MB/s. These findings indicated that the network did not have the available bandwidth for increased block sizes. Raw data from the test can be found at http://pastebin.com/raw.php?i=6b4NuiVQ. Peter suggested running tests on a simulated network with worst-case network failures to determine the necessary safety margin.In summary, these discussions and tests highlighted the challenges and limitations in increasing block sizes in the Bitcoin network due to bandwidth constraints. The need for careful evaluation and consideration of factors such as decentralization, security, and economic incentives was emphasized.
Updated on: 2023-08-01T14:29:57.477609+00:00