How tunnels work

An IPSec tunnel is typically thought of as a discrete two-way logical connection between two systems. However, each tunnel is actually made up of complementary unidirectional connections called Security Associations, or SAs.

Each SA is identified by a 32-bit number called the Security Parameter Index or SA ID.

SAs define the attributes of an IPSec tunnel, the most important of which is the security protocol being used.

The security protocol determines how data traffic which passes through the tunnel is to be encoded and decoded. There are two different protocols; either or both may be used in any given IPSec tunnel, depending on what form of security is required.

These security protocols operate by encoding every IP packet sent over the tunnel, using a predetermined algorithm.

Note: The DES-CBC algorithm is only available in MPTS V5.4 (WR_8610) and above. MPTS V5.3 (which came with TCP/IP V4.1) did not include the DES module due to U.S. Government export restrictions which were in effect at the time of release.

Transport vs. tunnel mode

With either protocol, there are two different ways in which IPSec may encode an IP packet.

Note that the term 'tunnel mode' has no relation to the term 'IPSec tunnel'. An IPSec tunnel may use either transport or tunnel mode.

When using tunnels in transport mode, you must disable the 'MTU Path Discovery' feature of the IP layer. This can be done with the command

    inetcfg -set mtudiscover 0


[Back: Definition of an IPSec tunnel]
[Next: Tunnel types]