*******************************

In the current Classfull standard, when you subnet, you wind up with 3 distinct IDs:

1) The original Network ID (defined by the default class mask)

2) The Subnet ID (defined by any bits beyond the default class mask)

3) The Host ID

For example, if you subnet a B block with 255.255.224.0, you get:

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0

\______________A_______________/\_B_/\___________C____________/

A: Network ID
B: Subnet ID
C: Host ID

This standard considers it to be 'illegal' (i.e. it strongly recommends against) to use any Subnet ID that is all 1's or all 0's (these values have special meaning to some devices. e.g. broadcast). This means that you 'must' always drop the first and last subnet.

You will find a lot of argument on this subject, because there is a new 'Draft Standard' called Classless Inter-Domain Routing (CIDR), which has been the common practice for a number of years. Under CIDR, you would be able to use all 8 subnets (you would still have to drop the first and last Host ID, though).

Most 'modern' equipment will support CIDR to one level or another (e.g. all MS stacks allow the use of the first and last subnets). However, you may run into old devices that won't support it.

MS's exam adheres to the current Classfull standard.

For more information, go to one of the Internic mirrors, such as:

FTP://RS.InterNIC.Net/rfc/

http://www.rfc-editor.org/

http://www.freesoft.org/CIE/RFC

and look up the relevant RFCs (e.g. 917, 950, 1517-20, 1818, 1878).

********************************