Path: utzoo!utgpu!water!watmath!clyde!cbosgd!ihnp4!ptsfa!ames!umd5!purdue!i.cc.purdue.edu!j.cc.purdue.edu!pur-ee!iuvax!bsu-cs!cfchiesa
From: cfchiesa@bsu-cs.UUCP (Christopher F. Chiesa)
Newsgroups: comp.os.vms
Subject: Re: PHONE command
Summary: Compromise?
Message-ID: <1869@bsu-cs.UUCP>
Date: 17 Jan 88 02:30:11 GMT
References: <880114112838.00000542.EAQI.83@WYOCDC1>
Organization: CS Dept, Ball St U, Muncie, Indiana
Lines: 37

I'm not a system manager, but have been a (student) user of the Ball State
University VaxCluster for the past two and a half years.  The cluster has 4
nodes (originally four 11/785's, now three 11/785's and an 8650), used by
several thousand students, faculty, and staff.  In the time I've been here,
I've seen a lot of genuine ABUSE of the PHONE utility; a conversation between
two or more people sitting a few terminals apart, chatting in PHONE and oc-
casionally commenting verbally ABOUT the PHONE conversation, isn't impossible,
for instance.  Some sort of control does seem to be needed, and our system
managers have tried several different measures.

The first control measure in my experience was the total removal of the PHONE
utility.  That measure released ALL resources used by PHONE -- but the
"gossipers" and "abusers" simply switched to using MAIL instead, thereby using
not only CPU and I/O resources, but now DISK SPACE also!  (Each user has a 
disk quota of 1000 blocks, but students' MAIL.MAI files grow to 900+ blocks
on a regular basis.)

Next (to relieve some of the MAIL abuse, I presume), PHONE was restored, amid
warnings that it would be "taken down again if abused" in the system login
messages.  Of course, students figured "Well, MY phone call won't hurt," and
abuse started up again.

Finally, after another brief PHONEless period, PHONE came back once more, but
subject to the restriction that it can now be used only between users
logged into the SAME NODE, rather than cross-node as had been possible previ-
ously.  This is where the situation stands today, and in my opinion it's a
reasonable compromise: cross-node I/O resources, at least, aren't used up by
PHONE, and a user trying to get work done only has to worry about interrup-
tion by users of ONE node, rather than FOUR.  (Does that constitute a quarter-
ing of the abuse potential, or am I being too simplistic?)  A lot of people
still converse through MAIL for hours on end, but at least not EVERYONE does.


Chris Chiesa
Senior, CS Dept.
Ball State University
Muncie, IN