ludey@umbio.MIAMI.EDU (Lanny Udey) (08/22/88)
We have several people here who need to mail to people on the JANET network. They have received various routings via UUCP, BITNET and the Internet but none seem to work consistantly for all sites. Is there a major gateway site off the Internet or a routing that you have found to be consistant? Any help would be appreciated. Please mail to me at: ludey%umbio@umigw.maimi.edu Internet or ....uunet!gould!umbio!ludey UUCP -- Lanny R. Udey Biomedical Computing
sjl@eagle.ukc.ac.uk (S.J.Leviseur) (08/23/88)
You should be able to mail the janet site without explicit routing over the arpanet. UK is a registered top-level domain. For example mail:- user@site.ac.uk The uucp gateway and the arpa gateway sites in the UK both run the same software so if one fails the other probably will as well... If you want to explicitly route mail from uucp -> janet then route it via ukc. However if the janet site is blacklisted by ukc the mail WILL be silently discarded. This is because ukc has to pay for all international traffic, both receiving AND originating so UK sites have to agree to pay their share. If they do not agree the ukc gate will just discard their traffic. This could be what is happening to your mail. postmaster@ukc
jim@cs.strath.ac.uk (Jim Reid) (08/23/88)
In article <319@umbio.MIAMI.EDU> ludey@umbio.MIAMI.EDU (Lanny Udey) writes: >We have several people here who need to mail to people on the JANET >network. They have received various routings via UUCP, BITNET and the >Internet but none seem to work consistantly for all sites. > >Is there a major gateway site off the Internet or a routing that you >have found to be consistant? There are three distinct gateways from the US into JANET - one for each of the main networks you mention. I'll describe each in turn. [1] INTERNET The JANET/Internet gateway is currently at University College London. It permits telnet and mail traffic to authorised users only. If either the sender or recipient of the mail message is not authorised, UCL won't forward the mail and they might not even return it with an explanatory message to you. Access to this gateway is restricted for reasons of cost and capacity. A new gateway is supposed to be installed "soon". This will be based at ULCC (University of London Computer Centre) and is supposed to provide free and unrestricted access for JANET users to the full range of Internet services. I understand that UCL have entered JANET domain names as MX records with the NIC. If your mail system is on the Internet and understands MX records, then mailing user@somesite.ac.uk will work if the user is registered with UCL. [The MX records point at the ARPA side of the UCL gateway.] Percent-style routing also works - i.e. user%somesite.ac.uk@ucl-cs.arpa - though this shouldn't be necessary on the Internet these days. [2] UUCP The University of Kent - UKC provide the UK's UUCP gateway. They too have restrictions on passing mail. The UUCP network costs real money and UKC recoup their costs from sites who register with them. For unregistered sites, UKC say that they will silently throw the mail away - sending it back costs them money. In practice, they let the mail through, but this will stop if the costs escalate. A bang-style path to a site somewhere in the UK would begin ....!uunet!mcvax!ukc!UKsite!user. If this site was registered with UKC, it should appear in the UUCP maps and would be pathaliased to user@UKsite.uucp. UKC can also accept domain names, so an address like uunet!mcvax!ukc!uk.ac.somesite!user would be OK, as would uunet!mcvax!ukc!somesite.ac.uk!user. [3] BITNET The JANET/Bitnet gateway is at Rutherford Laboratories. It is supported by IBM and is totally free until the end of the year when IBM will review their funding. In Bitnet terms, this gateway is called UKACRL. It uses percent-style routing, so an address like user%uk.ac.somesite@UKACRL.BITNET should work. However, the BITNET gateway is not very clever - it can't do things a UK gateway has to do (like flip domains around - more on this later). ADDRESSING FORMATS JANET uses a mail protocol that's more or less identical to the Internet standard RFC822. The important difference is in domain ordering. JANET puts the biggest domain at the left, the rest of the world puts it on the right! In JANET terms, my address is jim@uk.ac.strath.cs but for everywhere else it's jim@cs.strath.ac.uk ! Now the gateways at UKC and UCL are super-smart. They can flip domains around, handle partially qualified domain names and cope with mixed format addressing such as a!b@c. Provided your mail has something reasonable as an address, UCL and UKC will do the right thing with it. Both the UKC and UCL mail systems manage to hide the stupidity of JANET domain ordering from the rest of the world. On the other hand, the Bitnet gateway can't do any of that. They can only get by with full JANET names in JANET ordering in either the long or short form of JANET names. i.e. user%UK.AC.SOMESITE@UKACRL.BITNET or user%UK.AC.SOMESITE-UNIVERSITY@UKACRL.BITNET should work. Anything else probably won't. If you want to use the BITNET gateway to get mail into JANET, you must ensure that UKACRL is presented with a mail address as outlined above. If not, the mail probably won't be understood by the gateway. From personal experience, I've never had trouble with the UCL or UKC mail systems. The sheer volume of mail they deal with means that they have to get it right and be tolerant of mail protocol violations. I have never been able to send mail into BITNET via the gateway and get a reply through the same gateway. I hope this is useful. Jim -- ARPA: jim%cs.strath.ac.uk@ucl-cs.arpa, jim@cs.strath.ac.uk UUCP: jim@strath-cs.uucp, ...!uunet!mcvax!ukc!strath-cs!jim JANET: jim@uk.ac.strath.cs "JANET domain ordering is swapped around so's there'd be some use for rev(1)!"
igb@cs.bham.ac.uk (Ian G Batten <BattenIG>) (09/02/88)
In article <1207@stracs.cs.strath.ac.uk> jim@cs.strath.ac.uk writes: > The JANET/Internet gateway is currently at University College > London. It permits telnet and mail traffic to authorised users only. If > either the sender or recipient of the mail message is not authorised, > UCL won't forward the mail and they might not even return it with an > explanatory message to you. My experience is that unregistered users (ie me) do receive mail through nss.cs.ucl.ac.uk. I'm in the process of moving my mailing-list subscriptions to bitnet (see below) but if people use MX-capable mailers any routing I specify often gets ignored. > I understand that UCL have entered JANET domain names as MX > records with the NIC. If your mail system is on the Internet and > understands MX records, then mailing user@somesite.ac.uk will work if > the user is registered with UCL. [The MX records point at the ARPA side > of the UCL gateway.] Percent-style routing also works - i.e. > user%somesite.ac.uk@ucl-cs.arpa - though this shouldn't be necessary on > the Internet these days. Again, registration into Janet is not (currently) required. One would hope that if it was, sites with NO registered users (ie here) would have their MX records point at either uunet or a bitnet gateway. > and UKC recoup their costs from sites who register with them. For > unregistered sites, UKC say that they will silently throw the mail away > - sending it back costs them money. In practice, they let the mail > through, but this will stop if the costs escalate. I think Peter Collinson said their were sending a "taster" to the recipient: "We have received mail for you from Xyzzy@foo.bar subject Whatever and because you are unregistered we have dropped it on the floor". > I have never been able to send mail into BITNET via the gateway and get > a reply through the same gateway. No problem! I run all my mail that way. Quote your return address as BattenIG%cs.bham.ac.uk@cunyvm.cuny.edu (or mitvma.mit.edu, or some other Bitnet/Internet gateway) and life is great! It's a LOT faster than ucl or ukc --- I get mail from my Sun to Internet sites in around a minute and back in about the same. If I mis-address something and it bounces at, say, bco-multics.arpa (to quote a recent example) the longest step in the loop is the fine-minute batching of local smtp traffic on our ethernet.