![]() |
![]() |
|
You're here: Home |
David Conrad wrote:
On Jun 3, 2008, at 7:00 AM, Jeroen Massar wrote:It's a bit sad that we didn't take the opportunity to define a new API that would return 'network handles'SCTP & HIP. Please look at these.This still requires applications to be far too aware of what's underneath them. Sort of like requiring applications to know SCSI protocol commands or disk geometry in order to write out data.
No they don't. Just open a socket() of IPPROTO_SCTP and get going, just like TCP. You miss out on all kinds of cool features of SCTP like multiple streams in one SCTP connection, but who cares about that.
Another nice problem here is that for people who correct programmed their code using the getaddrinfo() loop, SCTP won't be returned when asking for STREAM sockets because getaddrinfo() doesn't support it. Then again, afaik there is no SCTP for MacOSX nor for Windows which is easily available thus that also blocks quite a bit.
Then again, quite some years ago there was no IPv6 anywhere, there is quite some of it now though, people just have to see the light(tm).
Now the big problem is, how do you get programmers to start adding SCTP and/or HIP support to all their programs...And how do you get CPE router vendors to support them and firewall administrators to allow them through, etc.
CPE "routers" should not care about it, if they are really routers they just route IP and nothing else. If you mean NAT boxes, well those should have died a horrible death ages ago, fortunately there is not supposed to be NAT in IPv6, people who NAT anyway should follow those NAT boxes.
Firewalls are indeed always a funny thing, they are the main culprit for not being able to do end2end as one would assume.
Greets, Jeroen
Attachment:
signature.asc
Description: OpenPGP digital signature