> Chris, > > I conceptually agree with the proposed change. > > At the same time, at the risk of duplicating prior comments, > I see this allows for the HOSTNAME to reference a logical > entity which may be a subset or superset of a physical > entity. For example, foo.bar.com may easily refer to a web > server process, a load-balancing switch for several e-mail > servers, or a vanilla-flavor physical host machine of > arbitrary use. This suggests that HOSTNAME does not and > cannot unambiguously (or authentically) identify the source > of the message. Is this a problem?
I don't think so - the current RFC3164 hostname has the same problem. Also, if you think about the front-ended web servers, the text does not require that you put the public name (e.g. www.example.net) into the hostname field - it could also be www-head.example.net, www1.example.net and www2.example.net if these are the three servers making up the farm. In fact, I would expect the later in, as these are what I think the real hostnames of the devices. I think the change Chris proposes makes life much easier - and is also already seen in many implementations. I need to re-read the RFC in regard to relays. In my view, we should allow a relay to reformat a hostname (any kind) in case it knows better about the real name. I am picking up on Andrew Ross' sample of many devices sending with a hostname of 10.0.0.1. A local relay could then be configured to rewrite the name to a more meaningful one before forwarding it to the central relay. Rainer > -----Original Message----- > From: Christopher Lonvick > > --vvv--- Proposed change to Section 2.2 ---vvv--- > > The HOSTNAME field will contain an indication of the originator of > the message in one of four formats: only the hostname, > the hostname > and domainname, the IPv4 address, or the IPv6 address. > The preferred > value is the hostname and domainname in the format specified in > STD-13 [RFC 1034 and 1035]. This format will be referred > to in this > document as HOSTNAME-STD13. If only the hostname is used, the > HOSTNAME field MUST contain the hostname only of the device as > specified in STD-13 [4]. This format is discouraged but > provides for > legacy compatability with the format described in RFC 3164. This > format will be referred to in this document as HOSTNAME-3164. In > this format, the Domain Name MUST NOT be included in the HOSTNAME > field. If the IPv4 address is used, it MUST be shown as the dotted > decimal notation as used in STD-13 [5], and will be referred to as > HOSTNAME-IPV4. If an IPv6 address is used, any valid > representation > used in RFC-2373 [6] MAY be used and will be referred to as > HOSTNAME-IPV6. A single space character MUST also follow the > HOSTNAME field. > > Messages containing Signature Blocks and Certificate Blocks as > described in this document MUST use the HOSTNAME-STD13 format in > the HOSTNAME field. > > ---^^^--- Proposed change to Section 2.2 ---^^^---
