[185] in Kerberos_V5_Development
timestamps in tickets.
daemon@ATHENA.MIT.EDU (Bill Sommerfeld)
Wed Feb 14 18:46:05 1990
Date: Wed, 14 Feb 90 15:39:11 EST
From: sommerfeld@APOLLO.COM (Bill Sommerfeld)
To: jtkohl@ATHENA.MIT.EDU
Cc: krbdev@ATHENA.MIT.EDU, jis@ATHENA.MIT.EDU, jon@ATHENA.MIT.EDU
I'm about to start beating a dying horse now..
I just saw a proposed spec of an "OSIfied" NTP done by someone at HP
Labs-Bristol; it came into our hands via someone at DEC.
The timestamp format they used was (almost) exactly the same as used
by NTP.
I quote from the document as follows:
"A TimeStamp is represented as a 64-bit unsigned ifxed-point number,
in seconds relative to zero hours on 1 January 1900. The integer part
is in the first 32 bits, and the frction part in the last 32 bits.
This format allows convenient multiple-precision arithmetic. The
precision of this representation is about 200 picoseconds, which
should be adequate for even the most exotic requirements.
TimeStamp ::= [APPLICATION 2] IMPLICIT SEQUENCE {
integer INTEGER,
fraction INTEGER
}"
This is much more convenient to deal with than the "standard"
timestamp format.
May I suggest we do the same in Kerberos?
- Bill
