[LEAPSECS] Bulletin C and all that

[Rob Seaman]

Zefram <zefram at fysh.org> wrote:
> Peter Vince wrote:
>> Class E addresses?
> 
> IPv4 address space is for some purposes divided into five classes:
> class A is 0.0.0.0/1, class B is 128.0.0.0/2, class C is 192.0.0.0/3,
> class D is 224.0.0.0/4, and class E is 240.0.0.0/4.  Classes A, B, and
> C form the unicast address space, originally with subnet sizes linked
> to the classes but now with no real class distinction.  Class D is the
> multicast address space.  Class E, apart from the limited-broadcast
> address 255.255.255.255, remains reserved for future use.

Richard Stevens used to work for the National Observatory (U.S.) back in the dawn time and the original edition of his TCP/IP Illustrated books used our class B network for examples.  It used to be a way to describe the size (or at least scope) of an organization whether they were class B or C.  (Very few class A.)  E.g., a large university might be class B, a university department class C.

Probably no risk attached, but we should perhaps exclude 255.255.255.255 from the encoding.  For my current encoding, PHK's CRC keeps the final octet out of harm's way:

	255.255.255.47 -> OK 2142 8 127 -1 (1, 1)

As shown, the first three 255s would require a negative leap second at the end of July 2142 with a specific prior offset.  It might be a bit pedantic to worry about this explicitly ;-)  Other encodings might be more likely but we can choose a CRC/SHA that avoids it.

Rob