# Is secure communication without public-key crytography feasible?

We all know about public-key cryptosystems, and we know some of these are computationally secure (for the time being, of course).

Are there any algorithms out there which allow safe communication between peers in a non “public/private” key fashion? If so, I would like to know some of these. If not, does any kind of formal demonstration of the infeasibility of such algorithms exist?

• Would you accept quantum crypto? – SEJPM Jun 19 '15 at 20:56
• Or simply symmetric crypto? I don't see anything in the question that rules that out. – Stephen Touset Jun 19 '15 at 20:57
• Would you accept TLS-PSK with TLS_PSK_WITH_AES_128_GCM_SHA256? RFC 5487 – SEJPM Jun 19 '15 at 20:59
• Something like TLS-PSK, but without pre-shared keys. I'm interested in the demonstration of infeasibility of secure communications with these constraints. – Alberto Fernández Jun 19 '15 at 21:04
• The whole point of public keys and signature is to verify identity of the parties and to agree on keys without trusted third parties. PKI allows the CA to sign keys offline and thus associate keys with entities. If you're willing to assume an online trusted third party, which shares a symmetric key with every participant, you indeed wouldn't need all this public key stuff. – SEJPM Jun 19 '15 at 21:18