jbvb@VAX.FTP.COM (James Van Bokkelen) (07/08/88)
If you send an ICMP echo request packet with IP Precedence = Flash Override and no IP TOS flag bits set, you back a reply with: Precedence = Routine from: vax.ftp.com Ultrix 2.0 gaak.lcs.mit.edu hacked 4.3bsd aim.rutgers.edu TWG VMS uunet.uu.net Sequent Balance sun.com SUNOS 4? Precedence = Flash Override from: xx.lcs.mit.edu TOPS-20 big-blue.lcs.mit.edu VM Wiscnet? ibm.com VM (IBM's 5798 FAL?) ucbvax.berkeley.edu 4.3bsd? Precedence = Flash Override & the Low Delay bit set from: oac.ucla.edu UCLA MVS James VanBokkelen FTP Software Inc.
Mills@UDEL.EDU (07/08/88)
James, Interesting. Try a nearby fuzzball (e.g. dcn1.arpa), which should return exactly what you send. Now, try all the critters with something TCPish and see what happens. Finally, slip in a source-route and see what comes back. Glad you brought this up. We never have really pinned down the specs in this area. Dave
jbvb@VAX.FTP.COM (James Van Bokkelen) (07/10/88)
dcn1.arpa returned the precedence I sent in a PING (which is widely perceived to be 'right', but is explicitly disallowed by RFC-792, pg 2). However, it does not escalate its precedence when I open a 'finger' TCP connection to it. (it does request High Throughput, though, which I didn't send). Recent discussion of TCP and Precedence reveals that RFC-793 and MIL-STD-1778 disagree as to whether an actively-opening TCP can raise its precedence. One authority I've talked to thinks the MIL-STD is right, and the opener can raise, but not everyone has been heard from. Sorry, I am not yet equipped to play with IP options, but I'll keep the fuzzies in mind when I start (Basic Security first). jbvb