kory@avatar.com (Kory Hamzeh) (05/10/91)
If I have a TCP/IP stack running on a MAC layer bridge to support SNMP, must the
ethernet interfaces have unique IP addresses?
What if the device was an IP bridge or router?
Thanks,
Kory
--
-------------------------------------------------------------------------------
Kory Hamzeh UUCP: avatar!kory or ..!uunet!avatar!kory
INTERNET: kory@avatar.com bierman@davidsys.com (05/14/91)
In article <1991May10.054726.2975@avatar.com>, kory@avatar.com (Kory Hamzeh) writes: > If I have a TCP/IP stack running on a MAC layer bridge to support SNMP, must the > ethernet interfaces have unique IP addresses? > > What if the device was an IP bridge or router? > > Thanks, > Kory I've been thinking about this problem also (Maybe you're also using the PCIP code ?) I hope this makes sense: * Each interface has its own MAC address. ==> one entry per interface in the ifTable * There is only network associated with the device (a 2-port bridge has an interface to two sides of the same network). There is not an 'IP address' for each interface. * The box may support one or more "entities" that have IP addresses. For an embedded SNMP Agent, this is usually just the IP address of the SNMP management entity. ==> one ipAddrEntry in the ipAddrTable for the SNMP entity This leaves us with the question: Q: What is the value of ipAdEntIfIndex for my one IP address? A: choose ONE of the N interfaces and always use that value I'm using the interface having the lowest MAC address (i.e. bridge port 1). I think as long as your SNMP Agent and ARP server agree on the IP to MAC address translation, there shouldn't be a problem. (This may not agree with RFC-1158, MIB-II) I think the problem is in the design of the ipAddrTable and the definition of ipAdEntIfIndex: "The index value which uniquely identifies the interface to which this entry is applicable. The interface identified by a particular value of this index is the same interface as identified by the same value of ifIndex." My solution may violate the *uniquely indentifies* clause. (It is up to the SNMP, ARP, and RARP implementations to enforce the uniqueness of the IP <--> MAC address mapping.) But--while there is a network (and maybe a sub-network) associated with each interface, (which may be the same for each interface), it doesn't make sense to say that an interface has an IP address, or that an IP address applies to a single interface. So I guess my answer (the short-form) is--NO, the IP address of each interface does not have to be unique. --andy; bierman@davidsys.com