Any disadvantages of using secondary IPs?

[Nukesquad]

I can't seem to figure out what is the practical difference and what are the consequences of using multiple secondary IP vs creating multiple logical interfaces?

I was wondering if this would bother my router's VoIP functions in anyway?

[globalnettech]

Hello,

problems, if any, with secondary addresses occur mainly in conjunction with routing protocols, such as OSPF, which cannot establish adjacencies on secondary addresses. Other than that, you can add as many as the software allows you to. I guess the reason why multiple secondary addresses are being used, rather than multiple logical interfaces, is that you do not have to bother with trunking, inter-VLAN routing etc. when using secondary addressing. It is just a simple way of being able to add more hosts. I would be careful, though, to not add to many hosts, since that, and especially with regard to VoIP, can cause problems with delay and dropped packets, which VoIP usually doesn't handle very well.

HTH,

GNT

[Richard Burts]

In addition to GNT's discussion of secondary addressing dependency in routing protocols I would add another aspect to be careful about when configuring secondary addressing: be careful that all routers on the segment agree which subnet is the primary address.

Also be aware that any packet sourced by the router from that interface will use the primary address as the source address and will not ever source anything from the secondary address.

My sense of the original post is that it was asking about the differences between configuring multiple secondary addresses and configuring multiple logical interfaces (which I assume means multiple subinterfaces). I believe the difference (and the choice of which one is appropriate) depends on the underlying topology of the network. Configuration of multiple secondary addresses is appropriate when the interface is into a single broadcast domain which would be a single VLAN or connected to an access port on a switch rather than to a trunk port. Configuration of multiple subinterfaces is appropriate when the are multiple broadcast domains which would be when connected to a trunk port of a switch.

HTH

Rick

[tdrais · Accepted solution]

From the routers viewpoint there is little difference since all broadcast traffic will be received by the router in both configurations The clients are a different isssue. With subinterfaces they only recieve broadcast from within that subnet.

If you were to put a IP phone on one subnet and pc on a different but put them on the same vlan they would appear to be separate but any broadcasts from the PC would still be received by the phones and cause issues. VoIP is much more sensitive to broadcast traffic than data.

In a worse case example if I were to run all my subnets as secondary addresses it would be almost the same as runnning huge subnet masks. But then there are people who still think that this flat earth concept is a good thing,

Secondary addresses are very useful in some backup situations but I would not design a network to use them when there are other options.

[scottmac]

IMHO, secondary addressing has one primary reason for existance: Orderly transition from one address block to another...then clients on both the old and the new addresses are concurrently operational. Once all the clients are migrated, the origianl address goes away and the new address is made primary.

Beyond that, any other use resluts in sub-optimal operation and performance.

All L2 broadcasts, from either/any secondary subnet, will be seen by all clients (L2 broadcast is all ones ... no distinction from the L3 address).

It doesn't give you any more capability for client count, it doesn't give you any additional capability for anything ... as mentioned, it can screw up your routing protocols and many consider it a security risk.

Using sub-interfaces is a fairly common way to address multiple PVCs on a frame-relay WAN, it's used with 802.1q encaps to speak with a trunk port on a switch (for VLAN routing / inter-VLAN routing).

There are still some adjustments to be made on how you implement your routing protocols.

Sub-interfaces have a couple useful specific roles, secondary addressing used for anything but address migration is a ticking time-bomb (IMHO).

FWIW

Scott