Re:HTTP/2.0 — Bad protocol, bad politics
by worklogin on Wednesday January 07, 2015 @03:50PM
(#8853225)
Yes, changing from TCP is about as difficult as changing from IPv4 - it would take years, and it would take a) catastrophic emergencies like IPv4's address exhaustion or b) completely passive, low-effort transitions/dual stack over years. Note that TCP processing is built into the hardware of all modern network ASICs from your PC to backbone routers and firewalls.
Unfortunate, but true.
Re:HTTP/2.0 — Bad protocol, bad politics
by gsnedders on Thursday January 08, 2015 @03:22AM
(#8855494)
What are the problems with SCTP over UDP, except the obvious extra (eight byte) overhead?
Re:HTTP/2.0 — Bad protocol, bad politics
by aidenn0 on Thursday January 08, 2015 @12:54PM
(#8857695)
UDP is just not as reliable over NAT (mainly due to crappy NAT implementations). SCTP tries really hard to keep it working (including heartbeats on idle connections), but implementing the 50% of SCTP that HTTP benefits from over TCP will work exactly as well as HTTPS whereas SCTP over UDP runs into lots of tiny issues due to nobody having tested it on their $5 NAT before putting it in a wifi router or a DSL/CABLE modem.
1) Unless tunneled over UDP (which has its own problems) failed to work with NATs and stateful firewalls.
2) HTTP(s) only environments (e.g. some big corporations) would not work with it; SPDY will look enough like HTTPS to fool most of these.
3) Lack of Windows and OS X support for SCTP (without installing a 3rd party driver) means tunneling over UDP.