On Fri, 01 Sep 1995 10:08:39 EDT, Jim Cocks said:
>  You need to send your request via an Internet/Bitnet gateway,
>  aka Interbit.  VPI has one and is near you so send a mail message:
>
>   Send to: [log in to unmask]
>   Subject:
>   SUB listname Julie Lopatofsky
 
Small correction 1: bitnic.bitnet is also reachable as bitnic.cren.net
- no gateway required.  Send your mail to [log in to unmask] and
be happy. ;)
 
Another small correction, please.. ;) We actually do gatewaying on
*two* machines here(*) but the "preferred" gateway for external use is
on vtbit.cc.vt.edu rather than vtvm1.  Also, the "canonical" gateway
that should be used unless you know of a better one for your location
is interbit.cren.net
 
Also, VPI is hardly "near" epix.net network-wise - a traceroute
reveals that (as of right now) packets are going from here via ANS to
Greensboro NC, to Atlanta, to Houston, to Albuquerque, to Los
Angelese, to San Francisco, throught the MAE-WEST interconnect, and
then another 5 hops on Sprintlink (which names their routers less
mnemonically than ANS) - 19 hops in all.
 
On the other hand, most of these hops are at T-3 speeds, and with the
current design of the long-haul carriers, I'd strongly suggest using
"service provided" rather than "distance" as the criteria for choosing
a gateway.
 
Bottom line - (1) first check if the Bitnet node you are interested in
in fact already has a working Internet connection, and use THAT
address if feasible.  (2) Unless you have a good reason not to, use
interbit.cren.net - at least you can be fairly sure *that* gateway
will be there as long as Bitnet is, whereas the vtbit gateway won't
survive past our departure from Bitnet(**).
 
                                Valdis Kletnieks
                                Computer Systems Engineer
                                Virginia Tech
 
(*) vtvm1 and vtbit share space on the same 3090-300 (no LPAR
involved, just a lot of lying with SYSTEM NETID files...)
 
(**) We are *NOT* contemplating leaving Bitnet at the current time.
However, there is a non-zero probability that we will not be the last
people out the door in 18 months or 2 years or whatever it takes for
Bitnet to finish dying.