Ref: 11640045
Title: Using a Subnet Mask
Date: 8/30/89

Copyright 3Com Corporation, 1991.  All rights reserved.

Subnet masks let you extend the bits used in an IP address
as the network number.  When using a subnet mask, remember these
requirements:

1.  All devices that are expected to communicate with each other
must have the same subnet mask, and the bits used for the subnet
mask must be set as on the other devices they wish to communicate
with.

2.  All binary 1's in the mask must match in the source address
those bit positions in the device receiving the packet.

3.  All binary 1's in the mask must be contiguous and the most
significant bit in the mask must be a binary 1.


For example, in a network with the Class A internet address
126.xxx.xxx.xxx, only the first field (126) is the network
number.  To divide this network into four subnets, the first two
bits in the second field of the IP address must take on meaning.
The appropriate subnet mask would be 255.192.000.000 (in
hexidecimal, FF.C0.00.00; in binary,
11111111.11000000.00000000.00000000).

A Communication Server using the subnet mask with IP address of
126.128.100.114 (in hexidecimal, 7E.80.64.72; in binary,
01111110.10000000.01100100.01110010) could communicate with a
device having an IP address which, in binary, would be
01111110.10hhhhhh.hhhhhhhh.hhhhhhhh (where an "h" can be a 0 or a
1).  The Comm Server could not communicate with a device having
an IP address in binary of 01111110.00hhhhhh.hhhhhhhh.hhhhhhhh,
because the first two bits in the second field of the IP
addresses do not match and are targeted by the subnet mask.

For more information on subnet masks, see the following sources:

* Gateway Server Installation and Operation Guide (09-0121-00),
Appendix G

* NCS/2 Installation and Operation Guide (09-0110-01), page 9-29

* Internetwork Bridge Installation and Operation Guide (09-0107-01),
Appendix E
