CCNA Making SUbnet Question

[Erik]

I am studying for the CCNA using the Cisco press book set. One of the questions in the end of the chapter on OSPF and EIGRP in the ICND book has me a bit confused.

Basically it says you have been assigned network 192.168.1.0, configure EIGRP on the router (one ethernet, two serial) with at most 60 hosts per subnet. So the ethernet needs to use a subnet mask of 255.255.255.192, and the serials need to use 255.255.255.252 since they are point to point. Easy enough so far. The e0 is configured with the ip address 192.168.1.65, the first available subnet address on that subnet.

What I don't really get is the serial is configured with 192.168.1.129. I realize that it is a valid subnet address on that subnet, but why use that when as far as I can see it would have been easier to just use 192.168.1.5? Is it just that the author felt like using that to show that you don't need to start with the first usable subnet address? Is there any real difference between choosing the two as far as the test is concerned? What really gets me is that the question says "When choosing IP address values and subnet numbers, you decide to start with the lowest numerical values."


Thanks for your help.

[deddard]

If you want to create a 30 bit subnet mask, it's usually best to chuck it in at the end of the range rather than halfway through - this gives you room for expansion later on. for example if you wanted 2 subnets with 60 hosts and one p to p link you could use (using the examples you have given):
0-63
64-127
and 128-131
but this would mess you up if you wanted to change the number of hosts later on. in real life a better way would be to use

0-63
128-191
248-251

The reason for this is that you have still got the room to expand both networks to hold twice the numer of hosts. other permutations would be to stick with the original 0-63, 64-127 but then also throw the point to point link in at the 248 or similar area. this means that you can add a furtehr 62 host network as well as a 30 host network.
THe original way (putting the p to p link in at 128) would prevent you from providing such flexibility.
The cisco books are a bit weird on this - I found a better way was to use the tables in tod lamles book (hope thats the right spelling)
you can create one yourself - create a table in word or (better still) a graphics package with 64 rows - at the very top (on the top line) write '0', on the next line down, write '4' and continue counting down in 4s til you reach the bottom (256)
now if you want to see what subnets you can use, just literally shade in the number of hosts on the table - you can start anywhere provided you follow the basic rules of:
You must follow standard binary rules - subnets can hold 4, 8,16,32,64 or 128 addresses
You must start at a point which is divisible by the number of hosts you need in total (e.g. if you want 60 hosts, you need a total of 64. therefore you can start this subnet at 0, 64, 128, or 192 - you couldn't start it at 8, 32 or some other number)

[Erik]

Thanks, it seemes a bit strange to me and the Cisco book didn't really explain it much. Hopefully on the actual exam the question is written so that either there is one only right answer, or they accept any technically correct answer.

I didn't read up on subnetting in the Lamles book yet, so I will check that out.