p2p: monoculture of DNS seeder software #16938

issue fanquake openend this issue on September 23, 2019
  1. fanquake commented at 4:45 am on September 23, 2019: member

    A comment from @TheBlueMatt in #15558:

    One high-level concern with this is the monoculture of DNS seed software we have, if you select three at random you’re almost certain to get three seeds serving from sipa’s seeder implementation, whereas we should really be trying to diversify a little bit (not to mention get things like dnssec going, which I dont believe sipa’s seeder support). Otherwise, Concept ACK.

    From @jonasschnelli:

    I think the risk of a single seeder codebase is orthogonal to this PR. Such risks could also be interpreted differently (diversity could also increase the risk based on the threat model).

  2. fanquake added the label Brainstorming on Sep 23, 2019
  3. fanquake added the label P2P on Sep 23, 2019
  4. naumenkogs commented at 9:18 am on September 23, 2019: member

    Somebody pointed me out to Grin’s DNS seeder in rust and told me that code quality is alright. Perhaps it would be easy for @TheBlueMatt to review it and fork, rather than building from scratch?

    Should we for once use a piece of software from altcoins? :)

  5. jonasschnelli commented at 12:08 pm on September 23, 2019: contributor

    From an architectural perspective, I think splitting the crawler from the dns server would be nice. One could use sipas crawler and use djbdns or bind or whatever to server the collected ips.

    IMO the server part has more risks for critical vulnerabilities, therefore reusing existing software for that part would be nice, although unclear if existing dns server implementations have the required flexibility to serve dynamic data.

    Unsure about the seeder from the Grin project. I couldn’t find crawling code at all (maybe I overlooked it).

  6. kristapsk commented at 9:29 pm on September 24, 2019: contributor

    From an architectural perspective, I think splitting the crawler from the dns server would be nice. One could use sipas crawler and use djbdns or bind or whatever to server the collected ips.

    This sounds like a good idea. I could look at bind for this (as this is something I have experience with from administrator perspective for 10+ years).

    unclear if existing dns server implementations have the required flexibility to serve dynamic data

    BIND has support for various database backends in addition to text configuration files, like LDAP, MySQL, PostgreSQL, Berkeley DB, ODBC, although it would be enough to just regularly update files and send SIGHUP to named.

  7. MarcoFalke commented at 11:38 am on August 15, 2022: member
    Not sure how this can be fixed. Seems an issue unrelated to the code in this repo
  8. MarcoFalke closed this on Aug 15, 2022

  9. bitcoin locked this on Aug 15, 2023

github-metadata-mirror

This is a metadata mirror of the GitHub repository bitcoin/bitcoin. This site is not affiliated with GitHub. Content is generated from a GitHub metadata backup.
generated: 2024-11-17 21:12 UTC

This site is hosted by @0xB10C
More mirrored repositories can be found on mirror.b10c.me