[BIP Proposal] Stale Tip Relay

The recent proposal for the BIP network was discussed, highlighting potential security vulnerabilities related to Denial of Service (DoS) and fingerprinting. The concern with DoS revolves around whether a small message could disproportionately affect processing, disk usage, or outgoing traffic. However, it is noted that sending invalid data as a DoS attack is not novel, as peers have several existing methods to achieve this. The proposed changes in the BIP do not introduce any new resource-exhaustion or bandwidth amplification vectors since processing constraints are limited by branch-length, recency, and retained-tip limits. Furthermore, nodes are not obligated to download outdated block data, and those that choose to do so will adhere to standard block-download mechanisms and resource limits. Implementations might enforce stricter controls, rate-limit abusive peers, or altogether disable problematic features.

Regarding the issue of fingerprinting, enabling staletip adds an observable characteristic, which could potentially be exploited. This addition is considered a trade-off for enabling the feature, which is typical for optional Peer-to-Peer (P2P) functionalities. Nodes that find this trade-off unfavorable have the option to disable the staletip relay. This functionality's implementation assumes distinct announcements for competing chains, where each message describes a single linear branch. For example, if A1-A2-A3-A4 represents the active chain, two competitive branches would be announced separately, identifying their fork points and subsequent headers.

These clarifications and responses aim to address concerns regarding the implementation and impacts of the proposed features within the BIP draft, ensuring that all parties understand the implications and operational details associated with these updates.