A critical mistake in Odom's INTRO and ICND book?

Sepiraph

I can't really believe this but it looks like either Odom made a rather large mistake by using the formula 2^n-2 for the number of subnet, whereas it seems like the convention from other sources indicate that CCNA would go with 2^n.

Odom's reasoning goes like this: the zero and broadcast subnet should not be used, so therefore you don't include the number. However based on my limited research this seems to go against the CCNA convention.

Netstudent

This is not a critical mistake on Odom's part. Maybe a mistake on the lack of research before you posted however. But don't worry we are all guilty of that from time to time. This has been talked about extensively in the forums. It is the difference between using the IP subnet-zero command and no ip subnet-zero. Check it out and do some research. The test will say wich to use in the question. Good luck!

tech-airman

Sepiraph wrote:

I can't really believe this but it looks like either Odom made a rather large mistake by using the formula 2^n-2 for the number of subnet, whereas it seems like the convention from other sources indicate that CCNA would go with 2^n.

Odom's reasoning goes like this: the zero and broadcast subnet should not be used, so therefore you don't include the number. However based on my limited research this seems to go against the CCNA convention.

Sepiraph,

This is a case of "both are right." The reason why they can both be right is based on other factors such as the routing protocol used. Since RIPv1 and IGRP are classful routing protocols, they follow the "2^n-2" rule. Since RIPv2, OSPF, EIGRP are classless routing protocols, they can use the "2^n" rule. The reason for the "zero and broadcast subnet should not be used" guideline is if a Cisco router on one end is connected to a non-Cisco router on the other side. The non-Cisco router on the other side may be using RIPv1 as it's routing protocol therefore may have problems with accepting routing updates for the zero subnetwork and the broadcast subnetwork in the RIPv1 routing table updates sent from the Cisco router to the non-Cisco router.

I hope this helps.

Source:

1. Subnet Zero and the All-Ones Subnet [IP Addressing Services] - Cisco Systems - http://www.cisco.com/en/US/tech/tk648/tk361/technologies_tech_note09186a0080093f18.shtml

Sepiraph

I'm aware of the difference between IP subnet-zero command and no ip subnet-zero.

Maybe I wasn't being clear enough, but in a multiple choice question that specifically ask for the # of subnetwork and host available per subnet if one apply a /# mask to a given Class A/B/C network, then it seems the convention that CCNA would go with is 2^n.

Now unless they specify somewhere on the CCNA exam (either available through online information or on the actual exam) that one should go with either, then confusion will arise. With your previous post, it seems that the test will indeed specify which convention.

Stotic

Todd Lammle, in his Sybex CCNA book, also uses the 2^n-2 format and mostly every single CCNA practice I take uses that approach.

Netstudent

You can also try www.subnettingquestions.com to sharpen up on the questions that explicitly specify using the Ip subnet-zero command.

Paul Boz

For all of the worrying that people do regarding ip subnet zero, it has never come up on a cisco test I've taken. I don't think arguing the semantics of that formula is really worth the effort or stress some people make it out to be.

Pash

Yeh, unless it explicity states in the questions that subnet zero is being used always assume you lose the network address and the broadcast.

Cheers and goodluck.