ejnorman@uwmacc.UUCP (Eric Norman) (01/07/86)
In article <383@ncr-sd.UUCP> greg@ncr-sd.UUCP (Greg Noel) apologizes: > traffic to/through this site to be dropped. Any traffic sent here > between January 2 (Thursday) and January 5 (Sunday) should be resent. Hypothetical situation: suppose I see this and conclude I needn't worry; I didn't send any mail through ncr-sd recently. But then in article <3080@sun.uucp>, chuq@sun.uucp (Chuq Von Rospach) informs us: > ihnp4 uses pathalias (or a variant of it) to do routing. So, how do I know ihnp4 didn't pathalias me through ncr-sd? Suppose further, that there was some way that I included "postage" with my mail to help ihnp4 pay its phone bills. Would I then be entitled to complain? If so, to whom? How long would it be before I even knew I had a complaint? Just wondering. -- Eric Norman UUCP: ...{allegra,ihnp4,seismo}!uwvax!uwmacc!ejnorman Pony Express: 1210 West Dayton Street, Madison, WI 53706 Life: Detroit!Alexandria!Omaha!Indianapolis!Madison!Hyde whoopie: yes date|touch|strip|join|sleep -- from ~Ewbanks/Makefile
greg@ncr-sd.UUCP (Greg Noel) (01/10/86)
In article <1878@uwmacc.UUCP> ejnorman@uwmacc.UUCP (Eric Norman) asks: > >So, how do I know ihnp4 didn't pathalias me through ncr-sd? Hmmmmmmm...... Interesting question, so I checked the path costs. It turns out that I have cheaper paths to four sites in San Diego than if they were routed through the local backbone (sdcsvax) (I didn't check other possible relay sites), and identical costs for seven more. In the latter case, the choice of relay is made arbitrarily (but consistantly) so I don't know which is being used. The bottom line is that if you sent mail to a San Diego site during the period in question, you should resend it. If it helps, ihnp4 called me with six pieces of traffic during those four days. I have the UUCP job numbers assigned by ihnp4, but I don't know if that can be used to trace the actual jobs. -- -- Greg Noel, NCR Rancho Bernardo Greg@ncr-sd.UUCP or Greg@nosc.ARPA
gjm@packard.UUCP (Gary J. Murakami) (01/10/86)
The current pathalias info on ihnp4 relating to ncr-sd is attached. This routing is all standard -- essentially what anyone running pathalias would get from the UUCP map data. All external sites are given a cost of HOURLY, so there isn't any rerouting to directly connected sites. Also FYI, here are the top neighbors for ihnp4 (neighbors that get the most unames routed through them): 5058 [all] 3613 [via] 233 seismo 216 okstate 199 akgua 170 allegra 154 uel 149 ucbvax 133 cbosgd 112 harvard 103 nsc 71 tektronix 67 topaz 67 hplabs ihnp4 delivers to 1445 (5058-3613) neighboring unames directly and routes 3613 unames through 205 of its neighbors. i.e. ihnp4 reachs 29% (1445/5058) of the UUCPMAP network directly! As for complaints about service, you are welcome to send me constructive critcism at <ihnp4!gjm>. However, there is little ground for complaint since ihnp4 is providing a relatively expensive service for free. -Gary ihnp4 pathalias output relating to ncr-sd: cost name route 500 ncr-sd ncr-sd!%s 500 ncr-tp ncr-sd!%s 500 sd.ncr ncr-sd!%s 700 onion ncr-sd!onion!%s 700 onion.ncr ncr-sd!onion!%s 700 ncrtps ncr-sd!ncrtps!%s 700 tranpro.sd.ncr ncr-sd!ncrtps!%s 700 pmncrrb ncr-sd!pmncrrb!%s 700 prodmgmt.sd.ncr ncr-sd!pmncrrb!%s 700 fiddler ncr-sd!fiddler!%s 700 saf.sd.ncr ncr-sd!fiddler!%s 700 se-sd ncr-sd!se-sd!%s 700 pld.sd.ncr ncr-sd!se-sd!%s 700 sdencore ncr-sd!sdencore!%s 700 man ncr-sd!man!%s 700 ziggy ncr-sd!ziggy!%s 700 ziggy.ncr ncr-sd!ziggy!%s 700 vrx.sd.ncr ncr-sd!ziggy!%s 795 jack ncr-sd!man!jack!%s 900 tw11 ncr-sd!se-sd!tw11!%s 900 tw11 ncr-sd!se-sd!tw11!%s 1000 isg100 ncr-sd!man!isg100!%s 1095 csm ncr-sd!man!jack!csm!%s 1400 rd1632 ncr-sd!rd1632!%s
greg@ncr-sd.UUCP (Greg Noel) (01/12/86)
In article <394@packard.UUCP> gjm@packard.UUCP (Gary J Murakami) offers
the ihnp4 pathalias output relating to ncr-sd.
Applying a little knowledge not directly available from pathalias, this
produces three sets of sites:
Our various names: ncr-sd, ncr-tp, and sd.ncr.
NCR sites for which we gateway: onion (onion.ncr), ncrtps (tranpro.sd.ncr),
pmncrrb (prodmgmt.sd.ncr), fiddler (saf.sd.ncr), se-sd (pld.sd.ncr),
ziggy (ziggy.ncr, vrx.sd.ncr), tw11 (why does tw11 appear twice?),
and rd1632. (Eight actual sites, despite the quantity of names.)
Other San Diego sites: sdencore, man, jack, isg100, and csm. (Five sites.)
I expected that the NCR sites would be there; with few exceptions, mail
for them must pass through us. What surprises me is the San Diego sites.
We offer the same costs to those San Diego sites also reached by sdcsvax;
I would have expected that either all of them would be listed or none of
them, so that if sdencore is there, all of the other San Diego sites should
be listed, too. (We do have a cheaper path to man (and thus jack, isg100,
and csm); I didn't realize that sdcsvax didn't speak to it directly.)
In any event, the bottom line is that if you had traffic for any of those
thirteen unlucky sites during the first few days of the year, you should
retry it.
At the risk of getting on a bandwagon, I, for one, would like to express
my appreciation to Gary, and to the current folks at ihnp4, for providing
us a marvelous service. They may occasionally get the details wrong, but
they can't always be faulted for that, and their hearts are in the right
place.
--
-- Greg Noel, NCR Rancho Bernardo Greg@ncr-sd.UUCP or Greg@nosc.ARPA