Re:HTTP 2.0
by justanother on Tuesday July 09, 2013 @08:39PM
(#6016667)
This was my main objection as well, after reading this. Make SPDY a standard, people say it's overengineered, but Goog and MSFT say it'll get us 10% performance benefit. Whatever, I trust them. And hey, think of all the next-generation tcpdump-style tools we'll get to write to decode the multiplexed connections (assuming any existing ones for SPDY won't already suffice).
But while you're going to the trouble to push a new version of HTTP, why not put some sort of mandatory encryption in there? Make the browser generate its own RSA key when it's installed? Require pinned self-signed certificates at the absolute minimum? Hell, I'm not an expert, I just want the NSA to have to work harder.
Encryption is only half the problem, authentication is the other. Self signed certificates don't authenticate the web server and allowing them would open the web up to MITM attacks. What we need is some form or decentralised certificate authority, i have no idea how that could work though.
HTTP allready supports LZMA, its just the browser and web servers that dont. If a browser sent the header:
Accept-Encoding: compress, gzip, lzma
instead of: Accept-Encoding: compress, gzip
and the server supported it, it would "just work". Unfortunately none of the major browsers or server have implemented it.Unencrypted HTTP is already open to MITM attacks. What I proposed was using self-signed certificates in conjunction with certificate pinning, which means that as long as your first connection to a website isn't MITMed, later connections cannot be MITMed or snooped on. As for the compression, I was referring to LZMA for header compression. I am well aware of implementation efforts of LZMA in Firefox and Chrome.
What we need is some form or decentralised certificate authority, i have no idea how that could work though.
DNSSEC could be used to distribute certificates: https://en.wikipedia.org/wiki/DNS-based_Authentication_of_Na...
Re:HTTP 2.0
by jimktrains2 on Tuesday July 09, 2013 @09:09PM
(#6016796)
Isn't DNSSEC still have centralized certs for the tlds?
We're also still stuck with DEFLATE, which is vastly inferior to modern compression formats like LZMA.