SCTP API draft (was Re: New "IP Version 6 Addressing Architecture" draft)

[La Monte Henry Piggy Yarroll]

The discussion below raises concerns with our proposed draft for
an SCTP sockets API.  SCTP is a reliable datagram transport developed
by SIGTRAN to carry telephony signaling information.

One distinctive feature of SCTP is direct support for multi-homed
hosts.  SCTP has the notion of a "primary address" to which it sends
most packets.  It may optionally send heartbeat messages to other
addresses of a correspondent node.  If the primary fails SCTP switches
to a different correspondent address.

The issue is that a single host may have both IPv4 AND IPv6
interfaces.  SCTP allows a single association (connection) to span
both kinds of interface.  In our API draft we have suggested af_inet6
as the socket type for associations with mixed address types.
Effectively, we can create a passive socket which listens for
connections on both af_inet and af_inet6 ports.

Jim Bound's comment appears to claim this is a bad idea:
>An af_inet6 socket should not accept a connection for an af_inet socket.  

May I invite you to read draft-stewart-sctpsocket-sigtran-00.txt and
comment?  Have we missed a critical IPv6 consideration?
-- 
Put no trust in extortion,              La Monte Henry Piggy Yarroll
set no vain hope in plunder;            NIC Handle LY
if riches increase,
do not set your heart upon them.

Jim Bound <[EMAIL PROTECTED]> writes:
>Mauro,
>
>>> but v4 mapped does not affect the ISV porting effort at all and in fact
>>> makes their job much easier if the platform they are porting to supports
>>> that paradigm.
>
>>you are absolutely right. my concern was about api issues. a modification
>>in the behaviour of af_inet6 passive socket, so that they are not allowed
>>to accept connections from af_inet sockets, would have imho nightmarish
>>effects. 
>
>An af_inet6 socket should not accept a connection for an af_inet socket.  
>Any implementation specific code path in a hybrid stack (different from
>a dual stack I think you know??) that does this or neglects the issue
>will have problems as you state.  If an implementation does this it is
>broken, the market will fix the correction.  Also we have never had this
>problem at any test event.  Also early on I have run ftp, telnet, etc at
>the same time and have not seen any bugs.  In fact what you ask is a
>test at present.  
>
>No where in any spec do we (the IETF or the XNET TBD API we are buildin
>the base for here on ipng group) require how one uses mapped,
>compatible, or native IPv6 addresses.  Nor should we other than to
>define them and make them available to applications via the API.
--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page:                      http://playground.sun.com/ipng
FTP archive:                      ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------