Carsten Bormann wrote:
Thanks for the new word "pain text password" :-)
A password including 4 or more different uppercase, lowercase, numeric and symbol characters, maybe. :-)
I’ve been in several places where people wanted to (and did) deprecate (parts of) a protocol because they didn’t like it for its (then!) primary use case. E.g., for IPv6, zeroed out UDP checksums were not taken over from IPv4 (†), because *at the time* people used them mostly for unsafe purposes (such as NFS without checksums, because of CPU constraints; I've lost data to this practice).
I'm afraid the primary use case for zero UDP checksum *WAS* for uncompressed realtime voice/image where a few bits of errors in a packet is better than entire loss of the packet. But, the use case became meaningless as Ethernet has its own CRCC and drop packets with *A* bit of error. Later, IPv6 even assumes all the datalink layers have strong enough CRCC. As such, NFS over Ethernet should be safe *UNLESS* bit errors occur at the source, destination or within relays at upper datalink layer or above. But the problem *WAS* well known as was documented by the E2E paper by Saltzer et. al. as: https://web.mit.edu/Saltzer/www/publications/endtoend/endtoend.pdf One gateway computer developed a transient error in which while copying data from an input to an output buffer a byte pair was interchanged, with a frequency of about one such interchange in every million bytes passed. As such, it was already crazy to turn off UDP checksum for NFS. As for IPv6, as rfc2460 already stated: The IPv6 version of ICMP [ICMPv6] includes the above pseudo-header in its checksum computation; this is a change from the IPv4 version of ICMP, which does not include a pseudo-header in its checksum. The reason for the change is to protect ICMP from misdelivery or corruption of those fields of the IPv6 header on which it depends, which, unlike IPv4, are not covered by an internet-layer checksum. UDP checksum was mandated to protect not payload but pseudo-header. As such, I can't find any point in rfc6936 not to protect the pseudo- header and think the rfc should be deprecated. Security of TELNET and SSH will be sent in a separate mail. Masataka Ohta
