FYS-MA@MAMMUTTI.UTU.FI (Matti Aarnio) (07/02/90)
Archive-name: nic.funet.fi/30-Jun-90
Original-posting-by: Matti Aarnio <FYS-MA@MAMMUTTI.UTU.FI>
Original-subject: FINTUVM in Memoriam
Archive-site: nic.funet.fi [128.214.6.100]
Reposted-by: emv@math.lsa.umich.edu (Edward Vielmetti)
Sic Transit Gloria Mundi
It is now 30th of June, 1990. 1h30m before midnight local time.
It is this much the FINTUVM system is left of its run time.
This system has had two names, depending what has been onlookers
own network. For BITNET we were FINTUVM, for Internet we were
MAMMUTTI.UTU.FI. Later name is a bit joking, Mammoth, a good name
for IBM 3033-U16 mainframe, which is quite huge in size, but not
too powerfull.
During our few years of existence in EARN/BITNET we got quite a
reputation by hosting quite popular file servers, and being first
site to have some sort of FTP server with plenty of remote sites
to choose from. (Rpiecs is limited to Simtel20 only, and BITFTP@PUCC
came along quite recently with different strategy.)
Our software for that server based on command analysis on Eric
Thomas' Revised LISTSERV, and at actual TRICKLE operation original
code was written by John Fisher at RPIECS.
LISTSERV allows easy addition of new 'external' commands and provided
us the system for parsing request addresses...
Fisher's TRICKLE contained all the tools to make functionality for
fetching files from remote systems and - via local cache for most
actively requested files - to send them to requesters.
Original system at RPIECS still is only for accessing Simtel20 files,
but ours had configurable set of external UNIX systems.
(Needed heavy modifications...)
Experiences with this system were quite educational...
At one time - last September ? - total outbound BITNET trafic from
finland was around 2GB (don't remember for sure), 55% for FINTUVM
and 40% for FINHUTC. The FINHUTC had then LISTSERV which made all
mailing list explodings for whole scandinavia...
You may guess what caused that huge flow...
Now our disappearence will cut short repertoites on these european
TRICKLEs which rely on our kind of primary servers... Half of material
was delivered from here.
However, all isn't lost though I did't save our TRICKLE to some
tapes. It has already several snapshots just waiting to be opened
somewhere and be put to operation. (With very propable recompilation
of my toolboxes...)
Still, I wish to forget this glorious single tasking system, and
just concentrate on building nice efficient UNIX version of it.
I have been doing it for quite some time already, but doing it
properly takes a lot effort...
Those fortunate who have Internet access, see what is at NIC.FUNET.FI.
(IP address: 128.214.6.100 - if your system can't use DNS systems.)
It is already connected to two BITNET nodes, but I don't know what
will be its official name... We shall see quite soon.
Now less than a hour to midnight, my last minute file transmissions have
gone out, most of (software) subsystems have been shut down.
I am ready to move this system to history.
Some time next monday IBM service will come along and take system apart,
later next week some company selling used computers will take it away.
University has already moved most of its computing users to a VMS system,
and for us UNIX patriots also provided nice Sun4/390 to play with.
(over 80% of users being on VAX doesn't harm me at all, that Sun has a lot
of kick for us very few using it :-)
If you had anybody here you wish to reach, ask from POSTMAST@FIRIEN
(that VAX) or postmaster@UTU.FI, or me.
If you want to know when this UNIX trickle is ready, hold your curiousity.
The more I get mail, the less I can code and debug... It will be announced
on RED-UG, and PCSERV-L, as well as several other lists.
/Matti Aarnio
Valid addresses: MEA@UTU.FI
MEA@FIRIEN.BITNET (not preferred)
MEA@NIC.FUNET.FI
21905::KONTU::MEA (SPAN/HEPNET)