home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!munnari.oz.au!yoyo.aarnet.edu.au!news.adelaide.edu.au!dstos3.dsto.gov.au!nichols
- From: nichols@dstos3.dsto.gov.au
- Newsgroups: comp.dcom.sys.cisco
- Subject: IP SUBNET ANSWERS
- Message-ID: <1992Dec19.115502.165405@dstos3.dsto.gov.au>
- Date: 19 Dec 92 11:55:02 CST
- Organization: Defence Science and Technology Organisation
- Lines: 59
-
- In article <527.829.uupcb@freddy.ersys.edmonton.ab.ca> murray.chapman@freddy.ersys.edmonton.ab.ca (Murray Chapman) writes:
-
- > I'm from the novell environment where a cable segment running IPX
- >protocol takes on a net number as an ID and I'm told this is not the
- >same when running IP. A call to cisco tech support told me that my first
- >problem was trying to do this:
-
- > Me 142.60.254.1 <-- Router --> 142.60.254.2
-
- > when what I really needed was:
-
- > Me 142.60.254.1 <-- Router --> 142.60.253.1
-
- > My question is, does this work when using multiple subnets on one
- >side of the router as follows or does the port on the router need an
- >address in each subnet that you want to route?
-
- > Me 142.60.254.1 <-- Router --> 142.60.253.1 - 254
- > 142.60.252.1 - 254
- > 142.60.001.1 - 254
-
- To answer the question one can consider that the situation can be interpreted
- two ways. So that either proposition can be true. The uncertainty is in what
- is meant by "multiple subnets on one side of the router".
-
- (Throughout this posting I assume that the network is subnetted with a
- ffffff00 mask. I infered that from the addresses in the example).
-
- That could mean a number of conventional IP subnets; each on it's own
- cable and seperated by a multiport router or series of routers. In which case
- the first proposition is true. Only the subnets which are directly connected
- to the router must be specified in the attatched interface's address. Any other
- subnets available through other routers will be learnt by the routing protocol.
-
- Alternately it could mean that you want to configure three IP subnets to
- exist on the same cable; so that all three are logically directly connected
- to the one port. This is only possible by the use of a special feature of the
- ciscos. They allow multiple addresses (which identify multiple subnets) on one
- interface; primary (normal) and additional single or multiple secondary
- addresses. All routers connected to this network must support this feature
- and their interfaces must also have an address in each of the secondary
- subnets. Otherwise more complicated counter measures to correct the routing
- would be needed.
-
- You can save yourself from unneccessarily complex problems if you forget this
- difference between IPX and cisco's IP and forget secondary addresses and
- multiple subnets; unless you must use them. The definition of an IP subnet
- is normally the same as an IPX network; except for this cisco feature.
- The cisco manual describes the intended applications of secondary addresses.
- They include - when you want more than 254 IP nodes on one physical subnet.
- - when migrating from an unsubnetted IP network to a subnetted one.
-
- We have used the feature for the later reason but are attempting to
- progressively remove all secondaries. We want to have a more conventional
- IP network. This will simplify debugging by removing uncertainty about
- interworking with non-cisco products, ie. we needed static routes to Apollo
- routers.
-
- Regards, Peter.
-