Hacker News new | ask | show | jobs
by pilif 5160 days ago
It is my understanding that it's very likely that HTTP 2.0 will be SPDY - maybe with some added extensions, but the basis will be SPDY.

A clear case where a working and apparently perfectly well backwards compatible vendor-specific implementation worked so well that it actually might get to be an official standard.

Now if only we could have SSL with name based virtual hosts or much wider use of IPv6 so that SPDY will actually be useful for a wide range of server administrators.
2 comments
kelnos 5160 days ago
Now if only we could have SSL with name based virtual hosts

We can: it's called SNI (http://en.wikipedia.org/wiki/Server_Name_Indication). Unfortunately browser support isn't quite there; recent versions the major desktop browsers do support it, but e.g. the stock Android browser does not (because Apache HttpClient does not).

link
aaronblohowiak 5160 days ago
We wouldn't need SSL with name-based virtual hosts if web browsers could use SRV records (and thus connect to different ports, so the server would know which cert to cough up without requiring the name.)
link
kelnos 5160 days ago
This doesn't really scale. Every SSL-protected vhost then needs its own port on the server. In current practice there are probably plenty of free ports, but it just seems like a poor choice overall.

(Regardless, as I mentioned above, name-based vhosts work just fine with HTTPS, using SNI. Much easier to fill in the gaps in browser support for SNI than to get everyone to adopt SRV records.)

link