[comp.dcom.sys.cisco] Multiple Interface Addresses

RRH@VM.NRC.CA (04/12/91)

J. Bashinski  mentions a technique for having multiple addresses on  one
port by using secondary options. i.e

    interface ethernet 0
    ip address 131.108.1.1  255.255.255.0
    ip address 131.108.2.1  255.255.255.0  secondary
    ip address 131.108.3.1  255.255.255.0  secondary

We have been using the following technique (cause at time there was no
secondary option).

    ip route 131.108.2.1 ethernet 0
    ip route 131.108.3.1 ethernet 0

It does the job  but which technique is better ?

Ratilal Haria

rv@deins.Informatik.Uni-Dortmund.DE (Ruediger Volk) (04/14/91)

In article <1991Apr12.121846.20016@nrcnet0.nrc.ca>, RRH@VM.NRC.CA writes:
|> J. Bashinski  mentions a technique for having multiple addresses on  one
|> port by using secondary options. i.e
|> 
|>     interface ethernet 0
|>     ip address 131.108.1.1  255.255.255.0
|>     ip address 131.108.2.1  255.255.255.0  secondary
|>     ip address 131.108.3.1  255.255.255.0  secondary
|> 
|> We have been using the following technique (cause at time there was no
|> secondary option).
|> 
|>     ip route 131.108.2.1 ethernet 0
|>     ip route 131.108.3.1 ethernet 0
|> 
|> It does the job  but which technique is better ?
The interface routes (the last alternative) are a quite old feature, so
I guess the secondary addresses are "better" :-)
In fact I have changed some interface routes to secondary addresses in my
configuration (though not on Ethernet but on X.25 interfaces); being able
to define an address for the router itself on the additional network is
useful - though it's still not perfect: the router does not use the
secondary address when sending packets out - which inhibits the use
of routing protocols on networks joined over secondary address assignments
(probably not so much a problem with LAN interfaces - but we are missing
this for our X.25 configuration; well, probably the best way to tackle
this specific problem could be using a different address binding 
architecture for the RFC877 encapsulation).

Ruediger Volk

rv@Informatik.Uni-Dortmund.DE