Article View: news.admin.technical
Article #6Re: Path: headers and small networks
From: Shuvam Misra
Date: Tue, 15 Sep 1992 14:19
Date: Tue, 15 Sep 1992 14:19
44 lines
2694 bytes
2694 bytes
In article <Bu7y4o.6s6@ux1.isu.edu> rhys@cs.uq.oz.au writes: | | Here's a technical question related to small networks in the outer regions | of USENET: is it a good idea to modify Path: headers as they pass out of | the small net into the world at large? [...] How exactly does the news software look at "Path:" headers and prevent loops in the feeding of articles? | not entering. Hopefully this will prevent another UUCP site with the name | "hosta" or "hostb" missing our news.... Why doesn't the news software just look at the first component of the path, and match that with the list of outgoing feeds? I know the obvious answer: that way, you don't get to prevent multi-hop loops. But even that can be stopped by adopting a related mechanism: when I get an article which I've already received, I don't re-accept it (thanks to the old records in the history file) and I _don't_forward_it_either_. This way, I break all duplicate article flow to me and through me, without anyone looking at the path header at all. What is the problem then? Seems to me that this approach is better than path-based filtering. Am I missing anything? Basically, it seems to me that there is duplication of information somewhere. Each site already has a history database of current and (slightly) old article ID's. This database is good enough to prevent all propagation of duplicates. Why do we need the additional path headers? One saving is of course, the transmission cost on the last leg of a multi-hop loop. Is that the _only_ reason to use this path based mechanism? And second question: is the Path header used at all for replying to an article? I thought the Sender and the From headers were used. (Hope this article is not inappropriate here; should I have posted to news.software.b instead????) Shuvam -- Shuvam Misra --------------------------- shuvam@jogin.cse.iitb.ernet.in -- -- Research Engr, Mach Project ---------------------------------------------- -- Dept of Computer Science, Indian Instt of Technology, Bombay 400076 ------ --------------------- "So much mail, so little time!" -----------------------
Technical Details
Message-ID:
<1992Sep15.113852.21122@cse.iitb.ernet.in>Path:
arc1-ca.pugleaf.net!sparky!uunet!sun-barr!ames!pacbell.com!tandem!zorch!scottReferences:
<Bu7y4o.6s6@ux1.isu.edu>Show Full Headers
Complete Article Headers
Newsgroups: news.admin.technical Path: arc1-ca.pugleaf.net!sparky!uunet!sun-barr!ames!pacbell.com!tandem!zorch!scott From: shuvam@cse.iitb.ernet.in (Shuvam Misra) Subject: Re: Path: headers and small networks Message-ID: <1992Sep15.113852.21122@cse.iitb.ernet.in> Keywords: path news admin Sender: scott@zorch.SF-Bay.ORG (Scott Hazen Mueller) Organization: Dept of Computer Science, IIT Bombay References: <Bu7y4o.6s6@ux1.isu.edu> Date: Tue, 15 Sep 1992 14:19:09 GMT Approved: scott@zorch.SF-Bay.ORG Lines: 44