[NTLUG:Discuss] network masks

[Rusty Haddock]

Cameron wrote:
    >Okay, after banging my head against the wall, I still don't have a firm
    >grasp on network masks like the following:
    >
    >  type              address/mask    netmask
    >  ----              ------------    -------
    >  Class B Private   172.16.0.0/12   255.15.0.0
    >  Class E Reserved  240.0.0.0/5     31.0.0.0
    >  IANA Reserved     96.0.0.0/4      15.0.0.0
    >  IANA Reserved     220.0.0.0/6     63.0.0.0

Your masks are wrong -- bits are "filled in" from the MSBit.
F'rinstance:

	Class B Private	172.16.0.0/12	255.240.0.0

	That /12 means the most significant 12-bits of the 32 in
	an IP address so:

		1111 1111 1111 0000 0000 0000 0000 0000

    		   255   .   240   .    0    .    0
	
	You were filling 'em in like this:

		1111 1111 0000 1111 0000 0000 0000 0000

		   255   .    15   .    0    .    0

	Class E Reserved  240.0.0.0/5     248.0.0.0
	IANA               96.0.0.0/4     240.0.0.0
	IANA              220.0.0.0/6     252.0.0.0

Here's that last one again on the video replay:


	MS bits	1234 5678 9012 3456 7890 1234 5678 9012
		1111 1100.0000 0000.0000 0000.0000 0000

MS bits are not "Micro$oft bits" -- no, we're talking about the
Most Significant Bits.

    >Could someone provide me with some examples of how a half-backed netmask
    >octet is handled (like 15.0.0.0 as opposed to the standard 255.0.0.0)?

I don't know all the rules but I don't believe that 15.0.0.0 is a valid
mask since the most significant bits are zero with one's following.

    >While we are on the subject, is there a good resource online that covers
    >how netmasking (or networking in general i suppose) works?  I read the
    >relevant sections of the LDP NAG, but it didn't go very deep into
    >netmasks nor routing depending on netmasks.  TIA

Two books that may help but they're relatively old -- still the information
remains the same after "all the years".

	"TCP/IP Network Administration", Craig Hunt, O'Reilly, 0-937175-82-x
	"Practical Internetworking with TCP/IP and UNIX",
		Smoot Carl-Mitchell & John S Quarterman,
		Addison-Wesley, 0-201-58629-0

    >--
    >cameron
    >  [ How do you tell when you're out of invisible ink? ]

No credible "spy" using invisible ink would EVER run out...

	-Rusty-
-- 
   _____        Rusty Haddock  =  KD4WLZ  =  rusty at fe2o3.lonestar.org
|\/   o \   o
|   (  -<  O o  Thanks, Sparky!          Charles M Schulz (1922-2000)
|/\__V__/