Published on: 2013-11-22T10:55:21+00:00
In an email exchange between Mike Hearn and Addy Yeow, Hearn praises Yeow's Bitnodes site and asks if it would be possible to gather and chart data on block and transaction propagation times. Yeow responds that there are several TODOs for the project in the next couple of months, including propagation data. Hearn also asks about the recent increase in running nodes and whether it is real or due to custom experimental additions. Yeow notes that a more in-depth analysis is needed.Hearn discusses additional logging he added to his node and notices that despite having 87 connections from regular nodes, virtually all transactions seen by his node are being announced by modified software running on different machines. He wonders if the owners of these nodes are relaying transactions without checking their validity, which may explain why they always win the race to announce to his node. Hearn has opened a pull request for his extra logging, which he considers trivial.On November 21, 2013, Hearn adds additional logging to his node and runs it for a few days. He finds that despite having 87 connections from regular nodes, virtually all transactions seen by his node are being announced by modified software running on different machines. Hearn wonders if anyone knows or owns these nodes and if they are relaying transactions without checking their validity, which seems to be the reason for their success. Jeff Garzik questions if the relay in question belongs to Matt and has reduced validity checking. The email exchange ends with an advertisement for Intel(R) Software Adrenaline and a link to sign up for it.Hearn shares a log showing the transactions accepted by his node and suspects that the modified software is relaying transactions without checking their validity, allowing them to always be the first to announce to his node. Garzik, a Bitcoin core developer, responds to this email. In another email conversation between Hearn and Arthur, Arthur shares a link to blockchain.info/hub-nodes, which shows three nodes from nogleg that relay the most to blockchain.info. Hearn questions if anyone knows or owns these nodes and if they are relaying transactions without checking their validity, which may explain their ability to win the race.The email thread also includes discussions on the recent increase in running nodes and the possibility of gathering and charting data on block and transaction propagation times. Hearn adds additional logging to his node and notices that virtually all transactions seen by his node are being announced by modified software running on different machines. He questions if anyone knows or owns these nodes and if they are relaying transactions without checking their validity, as it seems to be the reason for their success. The email provides a link to bitcointalk.org for further discussion.Hearn sends an email to the Bitcoin development mailing list on November 22, 2013. He adds additional logging to his node and runs it for a few days. He observes that virtually all transactions seen by his node are being announced by modified software running on different machines. Hearn wonders if anyone knows or owns these nodes and if they are relaying transactions without checking their validity. He shares a link to the pull request for his extra logging, which is considered trivial.In summary, Mike Hearn, a Bitcoin developer, discovers that despite having multiple connections, virtually all transactions seen by his node are being announced by modified software running on different machines. He suspects that these nodes are relaying transactions without checking their validity, allowing them to always be the first to announce to his node. Discussions on the recent increase in running nodes and the possibility of gathering data on block and transaction propagation times are also included in the email exchange.
Updated on: 2023-08-01T06:42:48.761290+00:00