info-vax@ucbvax.ARPA (08/21/85)
From: winalski%speedy.DEC@decwrl.ARPA Message-id: 19850820141026/5290@BABEL Posted-date: 20-Aug-1985 @ 14:10:26 To: rhea~decwrl~"OC.GARLAND@CU20B.ARPA" AT MAIL-11 Cc: rhea~decwrl~"INFO-VAX@SRI-KL.ARPA" AT MAIL-11 Reply-to: Paul Winalski AT ELUDOM Precedence: N References: INFO-VAX submission concerning Hardcopy article From: PAUL WINALSKI The article you read has it wrong. LAT-11 data packets containing terminal to host session data are not multicast. They are sent to the host's specific address. Now, it IS true that a single LAT->host message packet may contain data for multiple terminal sessions on that host. This should not present a security problem, though, if you trust the LTDRIVER on the host to correctly route the bytes to the proper terminal session. LAT-11 does use multicast, but only for circuit initiation. When a LAT server box comes on-line, it multicasts to all hosts to say it's available. When a host comes on-line, it multicasts to all servers to say it's available and which service classes it accepts. --PSW
info-vax@ucbvax.ARPA (08/21/85)
From: SYSTEM%CRNLNS.BITNET@WISCVM.ARPA Paul, Well, that's a relief. At least our LAT-11s aren't a load on the hosts. But what about DECserver-100s? Given that they are implemented using 68000s with all new code, would the LAT protocol preclude THEM from broadcasting their terminal packets? Selden Ball Cornell University, LNS NYNEX: (607) 256-4882 Wilson Synchrotron Lab BITNET: SYSTEM@CRNLNS Dryden Road ARPA: SYSTEM%CRNLNS.BITNET@WISCVM.ARPA Ithaca, NY 14853