matt@ncr-sd.UUCP (01/28/87)
In article <613@cdx39.UUCP> jc@cdx39.UUCP (John Chambers) writes: >A related topic that I haven't heard discussed: We have some uucp >links that are internal to the company, and are rather expensive >long-distance calls. For company business, it is often a good >idea to encourage their use. It's easier to uucp a big chunk of >source code across a phone line than it is to mess with tapes or >floppies, and it gets the job done much faster. But management >has some reasonable concerns that the outside world start using >these links for mail. We (ncr-sd.UUCP) have had this problem for a while because of incorrect costs associated with links over which we have no control. From our mod.map entry: # # Long-distance calls. Ihnp4 is the AT&T mail hub in Indian Hill (near # Chicago) that talks to just about everybody, and pyramid is in Palo Alto. # These calls are more expensive, so we bias them accordingly. Perhaps they # should be EVENING, but sdcsvax lists all of its long-distance connections # as DAILY and will cheerfully route everything from the San Diego area via # us if we don't take precautionary measures. # ncr-sd ihnp4(DAILY), pyramid(DAILY), cbosgd(DAILY), cbatt(DAILY) The fix is actually quite easy. We have local map information that is needed to describe our local domain as well as private links. Pathalias will use the cheapest cost given for any link so all we have to do is include the real cost in our local map. The values that we use are: # The published information puts a high cost on these links so that other # sites avoid routing through us. This overrides that information so that # we can use the links ourselves..... The costs are carefully constructed # so that we chose cbosgd in preference to ihnp4 whenever possible. (They # talk to each other with a cost of 95(DEDICATED).) ncr-sd ihnp4(HOURLY), cbosgd(HOURLY-94), cbatt(HOURLY-94-24), pyramid(DEMAND) I'm in the process of creating a company-wide newsgroup so this information can be automatically distributed. Information within a domain can easily be hidden by restricting the Distribution: to just that domain. This also allows us to publish changed information immediately, so internally at least, our routing information is always up to date. -- Matt Costello, matt.costello@SanDiego.NCR.COM (registered w/ CSNET) {sdcsvax,cbatt,dcdwest,nosc.ARPA,ihnp4}!ncr-sd!matt
brian@sdcsvax.UUCP (01/28/87)
In article <1323@ncr-sd.UUCP> matt@ncr-sd.UUCP (Matt Costello) writes: >We (ncr-sd.UUCP) have had this problem for a while because of incorrect >costs associated with links over which we have no control. From our >mod.map entry: > ># ># Long-distance calls. Ihnp4 is the AT&T mail hub in Indian Hill (near ># Chicago) that talks to just about everybody, and pyramid is in Palo Alto. ># These calls are more expensive, so we bias them accordingly. Perhaps they ># should be EVENING, but sdcsvax lists all of its long-distance connections ># as DAILY and will cheerfully route everything from the San Diego area via ># us if we don't take precautionary measures. ># Actually, they ARE the correct costs. Sdcsvax doesn't place long-distance calls at all [for political reasons that are (regrettably) inargueable and complicated - sigh]. We will hold mail for hosts such as ihnp4 and the like, but ITS UP TO THEM TO POLL US We have made our map entry (reproduced below) as reflective of our real-world connectivity as is possible; for the hosts that we do call, we clearly indicate that, and for those who call us, I have used the 3-month average of poll frequency as shown by our uucp records. As a reminder, DEDICATED means highspeed direct connection, DIRECT means a local call placed on demand, DEMAND is a long-distance call placed on demand (we use it for internet and X.25 connections), EVENING is a time restricted call (in this case, the only toll call that sdcsvax does make - to ncr-sd, the people who were complaining above), and all the rest we show are incoming-poll-only connections - they call us, we don't call them. In many cases, this takes advantage of these hosts' WATS or extended-area dialing coverage that we simply don't have. This, to my mind, is a whole lot closer to reality than most of the hosts out there that claim free connections to the entire net. Our map entry: sdcsvax ucsdhub(DEDICATED), apunix(DIRECT), bang(DIRECT), bigbang(DIRECT), calmasd(DIRECT), celerity(DIRECT), csndvax(DIRECT), dcdwest(DIRECT), esosun(DIRECT), gould9(DIRECT), hp-sdd(DIRECT), jack(DIRECT), man(DIRECT), net1(DIRECT), nosc(DIRECT), pdsfac(DIRECT), phonlab(DIRECT), scubed(DIRECT), sdsu(DIRECT), talaris(DIRECT), vis(DIRECT), wolf(DIRECT), ucbvax(DEMAND), ucsbcsl(DEMAND), rutgers(DEMAND), telesoft(HOURLY/2), piaget(HOURLY*2), sdencore(HOURLY*2), ncr-sd(EVENING), alcyon(DAILY), caiac(DAILY), ucdavis(DAILY), uw-beaver(DAILY), clyde(DAILY/5), bonnie(DAILY/4), caip(DAILY/4), scs-sd(DAILY/4), crash(DAILY/3), ihnp4(DAILY/3), moss(DAILY/3), ucrmath(DAILY/3), akgua(DAILY/2), cbosgd(DAILY/2), floyd(DAILY/2), gatech(DAILY/2), loral(DAILY/2), allegra(DAILY*2), bellcore(DAILY*3), decvax(WEEKLY), noao(WEEKLY), philabs(WEEKLY), scgvaxd(WEEKLY), sdcrdcf(WEEKLY), topaz(WEEKLY) --- Brian Kantor UCSD Office of Academic Computing Academic Network Operations Group UCSD B-028, La Jolla, CA 92093 USA