Talk:Key server (cryptographic)
|WikiProject Computer Security / Computing|
|WikiProject Cryptography / Computer science|
List of keyservers?
Wouldn't this be an appropriate page to list some indices of popular keyservers? -- anon
- Perhaps, although we want to avoid adding lots because Wikipedia discourages large directories of external links. One idea would be to link to a page which lists lots of key servers. — Matt Crypto 07:13, 11 Mar 2005 (UTC)
- Matt, could you make this a redirect to Key server (cryptographic), I made a disambiguation and started expanding the article. But, I don't think I have access privileges to make a redirect. I'm not 100% sure I did the disambiguation correctly, so feel free to make what ever changes. Thanks. — V. Alex Brennen Fri Nov 4 10:21:57 EST 2005
- PGP Corp. has created a centralized keyserver due to the difficulty of implementing the openpgp protocol. Many existing keyservers have difficulty handling subkey packets (damaging them), and ignore many other important types of packets altogether. Therefor, people are probably better to run with their default configurations for now rather than making use of listings. — V. Alex Brennen Thu Sep 29 12:41:09 EDT 2005
- Righty, that's done. — Matt Crypto 19:04, 4 November 2005 (UTC)
Web of trust
This article really doesn't deal much with the role of the public key server in the web of trust. Some mention of the fact that many public key servers enable other users to update your public key with their trust certifications (signatures) would be useful, as this is one of the advantages of this over other distribution methods. It would also bear mentioning that many of these servers communicate information about key revocation.
Some mention of synchronization between various public key servers and its consequences would also be useful.
Barring negative reaction, I would happily make these changes.
Reskusic 19:20, 8 March 2007 (UTC)
- I think this would be a wonderful addition. Key Revocation is one aspect that I think is really missing in the article. I think this is needed as this is how the "plaque" problem is supposed to be addressed. Old signatures (pre-revocation) would then still be treated as valid and the problem of bogus keys would be averted.
This section is really confusing in my opinion. It looks like PGP is somehow less secure than the whole PKI? There's no difference between publishing certificates/keys which are public. This sounds like a typical security by obscurity, and it's sounds weird. Every security system can be broken and PGP is not an exception. If the attacker cannot obtain your public key, means nobody can means there's no way for the people in the Net to tell who you are, means PGP/PKI is useless - anonymous. —Preceding unsigned comment added by 126.96.36.199 (talk) 09:38, 25 September 2007 (UTC)
Problem with "Problems with keyservers"
This section seems to say that "PGP Global Directory" (which I admit I have never heard of) solves the problem of not proving ownership of the keys (and indirectly that this is a problem). Is this a widely shared view? People who use PGP typically require a photo ID to prove ownership — that's significantly stronger than having control over an e-mail account for a few minutes. JöG (talk) 21:40, 9 December 2009 (UTC)
Update: PGP Corp also discusses this issue at https://keyserver.pgp.com/vkd/VKDVerificationPGPCom.html. Looks sane to me. JöG (talk) 21:45, 9 December 2009 (UTC)
The stale key problem can be solved by requiring an expiration date for a key and the auto-purge of expired keys (after a grace period). If someone wants to set their public key expiration date to be more than 10 years in the future (e.g. year 2038), that's their choice -- but it will eventually expire and be removed. 188.8.131.52 (talk) 23:29, 17 November 2010 (UTC)
I think an important point is that since the pgp global dir system is only using email verification anybody who has/gets access to the email account can delete a key and upload a new one (or might also have access to the private key if they've cracked a users account and have access to both e-mail and private key). This seems wildly insecure to me. I certainly won't trust any keys from pgp global dir, and I think this large security hole should be pointed out in this section. — Preceding unsigned comment added by 184.108.40.206 (talk • contribs) 22:54, 21 January 2012