Chronological Thread 

< Chronological >      < Thread >

---

Dear all,
see my comments in-line.
Best regards
Michael
On Wednesday, May 29, 2002, at 11:42 AM, Hideki Kishida wrote:

Hi,

Is not it using destination address that received (local
addresses) it in the logic of the association distinction of
SCTPLIB ?
Although it is testing with the following connection pattern it
has failed to the establishment of the SCTP association.

INIT
--------->
+----------+ Asc.1 +-----------+
| UA1 IP 1 | <--------> | IP 0 UA1|
| | Asc.2 | |
| UA2 IP 2 | <--------> | IP 0 UA2 |
+----------+ +-----------+
same port number same port number
UA1=UA2=4555 UA1=UA2=5555

Asc.1 Established (first)
Asc.2 Failed to establishment (second)

The problem is the following: the sctplib does NOT control the source
address of the SCTP packets. That is done by the IP stack. Because
the INIT is send to the same destination address, the source address
is the same (based on the frowarding table of the left host). It seems
that the IP stack always chooses IP1 as the source. That is why assoc 2
is not coming up.

I think that it is described that it should identify the
association possibility, by judging the address of UA1 and UA2
(as destination or local address), the above connection pattern
in RFC 2960.

You are correct, the above scenario is in the scope of RFC 2960 and should
work. But the SCTP implementation needs to do source address selection.
This is what the sctplib does not and will not in the (near) future.
Everything will work if you use also two different IP addresses on the right host
and get the forwarding set up correctly. Example
IP1 = 192.168.1.1/24
IP2 = 192.168.2.1/24

IP3 = 192.168.1.2/24
IP4 = 192.168.2.2/24
That will work.

-- RFC 2960 ----------
3.1 SCTP Common Header Field Descriptions

Source Port Number: 16 bits (unsigned integer)

This is the SCTP sender's port number. It can be used by the
receiver in combination with the source IP address, the SCTP
destination port and possibly the destination IP address to
identify the association to which this packet belongs.
-----------------------

TIA,
Hideki

--
Hideki Kishida
<kishida AT secnet.co.jp>
(Japan)

_______________________________________________
discussion mailing list
discussion AT sctp.de
http://www.sctp.de/mailman/listinfo/discussion

Michael.Tuexen AT micmac.franken.de

---

• [SCTP-Discussion] identification of association, Hideki Kishida, 05/29/2002
  • Re: [SCTP-Discussion] identification of association, Michael Tuexen, 05/31/2002

---

Archive powered by MHonArc 2.6.19+.