Posted by Eric Voskuil
May 16, 2026/22:39 UTC
The recent discussion in the Bitcoin Development Mailing List addresses common misconceptions about peer-to-peer (P2P) download speeds within the network. Notably, a key point highlighted is that P2P download speed is predominantly determined by the user's own bandwidth rather than the bandwidth of other peers. If a user has surplus bandwidth capacity, they can enhance their download efficiency by connecting to more or faster peers, and disconnecting from the slower ones. This method leverages the use of standard deviation to identify and drop slower peers, thus optimizing bandwidth utilization. However, it was mentioned that certain implementation issues in node software might prevent this ideal behavior, such as persistent connections to slow nodes until a timeout occurs.
Moreover, the explanation clarifies how the Bitcoin protocol manages connections with different types of nodes. Nodes advertise their capabilities via service bits, which should prevent connections to pruned peers for downloading blocks. This design intends to streamline network operations and maintain efficiency, emphasizing that pruning may adversely affect network performance. Over twelve years of experience in writing node software has shown that peers do not impose limits on download capacities, and the network is capable of reaching download speeds that match the theoretical limits dictated by a user's bandwidth.
Additionally, the email points out that throttling or reduced download speeds are generally due to hardware limitations or software design rather than restrictions imposed by peers. It also references a decline in relative download times over the past decade despite increasing data demands, suggesting improvements in network efficiency and bandwidth management. For further understanding, a link to a list of countries by internet connection speeds provides context on global bandwidth availability (List of countries by Internet connection speeds). The comparison to very low-speed internet services in certain countries underscores that the issues discussed are not limited by peer capabilities but possibly by other technical constraints.
Thread Summary (14 replies)
May 5 - May 16, 2026
15 messages • 14 replies
TLDR
We’ll email you summaries of the latest discussions from high signal bitcoin sources, like bitcoin-dev, lightning-dev, and Delving Bitcoin.
We'd love to hear your feedback on this project.
Give Feedback