0% found this document useful (0 votes)
3K views2 pages

SCTP Multi-Homing Explained

SCTP multi-homing allows a host to be addressed by multiple IP addresses and utilize multiple network interfaces to improve network resilience. It uses two paths per association, with data normally sent on the primary path and automatically retransmitted on the secondary path if the primary fails. SCTP monitors the secondary path and retransmits unacknowledged data through the secondary path if the primary fails, without affecting the application. There are symmetric and asymmetric types of SCTP multi-homing configurations to provide redundancy between peer networks.

Uploaded by

Sumit Tiwari
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
3K views2 pages

SCTP Multi-Homing Explained

SCTP multi-homing allows a host to be addressed by multiple IP addresses and utilize multiple network interfaces to improve network resilience. It uses two paths per association, with data normally sent on the primary path and automatically retransmitted on the secondary path if the primary fails. SCTP monitors the secondary path and retransmits unacknowledged data through the secondary path if the primary fails, without affecting the application. There are symmetric and asymmetric types of SCTP multi-homing configurations to provide redundancy between peer networks.

Uploaded by

Sumit Tiwari
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOC, PDF, TXT or read online on Scribd

SCTP multi-homing

In the IP terminology, a host is called multi-homed if it can be addressed by multiple IP addresses. To


make full use of the SCTP multi-homing, the host also needs to support multiple network interfaces,
each of which has to be configured to work in a different sub-network. The SCTP multi-homing is
used only for recovering from network failures. It is not used, for example, for load balancing. The
SCTP implementation supports two paths per association. Normally data is sent through the primary
path. If a network failure occurs in the primary network, SCTP resends unacknowledged data through
the secondary path automatically. The application cannot see which path is used and it does not affect
sending of data traffic in application level either. In this case, the SCTP stack takes care of all the
details.

The SCTP association works normally so that data runs through the primary path and SCTP heartbeat
runs through the secondary path. If something unexpected happens in the primary path, the SCTP
usually has an alternative path available. The SCTP monitors a condition of the secondary path all the
time by using a heartbeat message. The following figure describes how the SCTP retransmission works
when both the primary and the secondary paths failed. The retransmission procedure starts always the
same way; the first retransmission is done through the secondary path. The example above is based on
the following SCTP parameters:

• Retransmission time-out, [Link] 300 ms


• [Link] 500 ms
• [Link] 2
• [Link] 4

For more information about SCTP parameters, refer to chapter Modifying SCTP association level
parameters.

Figure: SCTP retransmission when both primary and secondary paths fail

The SCTP sends data first through the primary path. It starts the retransmission timer and starts to wait
for the SACK message from the remote peer. If the retransmission timer expires, the SCTP re-sends
data through the secondary path and restarts the retransmission timer again by using the same value as
before. If the remote peer does not send any SACK before the retransmission timer expires again, the
sending peer re-sends data packets through the primary path and doubled the value of the
retransmission timer. If new retransmission timer value is greater than the [Link] value, the SCTP
starts using the value of the [Link] parameter as a new retransmission timer. If both paths are totally
out of order, the remote peer does not send any SACK. Sending peers try to re-send data as many times
as they are allowed to try. The maximum amount of retransmission per association is the same as the
value of the [Link] parameter. When error counter is equal to or greater than the value of
[Link], the sending peer stops trying to resend data and closes an association by sending an
ABORT message to remote peer.

There are two types of SCTP multi-homing.


Figure: Symmetric network layout

The first type is symmetric SCTP multi-homing where both peers have two or more external Ethernet
interfaces available and connection is made between them. Each signalling unit contains two Ethernet
Interfaces in network element. The SCTP can use both interfaces in such a way that one is working as
a primary and the other one as a secondary path. The SCTP ensures that none of the messages can get
lost if only one path is broken at the same time. Each Ethernet port should belong to a different
subnetwork of the signalling unit, otherwise multi-homed connections do not work as expected.

Figure: Asymmetric network layout

The second type is asymmetric multi-homing where a configuration also has two separate paths but
Host A has only one Ethernet port in use. If the only Ethernet port of Host A or the gateway router
where it is connected fails, then multi-homing does not help. If Host A is assigned to a second network,
then the resiliency of the network can be improved. This additional address of Host A should be taken
into account in the network configuration of the routers as well. As a result of this modification, in case
of any kind of network failure, multi-homing helps to avoid association loss.

Description of how asymmetric SCTP multi-homing configuration should be configured is described


in the chapter Creation of SCTP configuration.

You might also like