|
|
| Author |
Message |
| 25 new of 306 responses total. |
davel
|
|
response 250 of 306:
| 
Jun 8 11:35 UTC 1998 |
However, last I heard that was *not* a Grex system proplem.
|
rcurl
|
|
response 251 of 306:
|
Jun 8 17:04 UTC 1998 |
You think "dummies" and "idiots" are not Grex system problems?
|
dpc
|
|
response 252 of 306:
|
Jun 9 02:09 UTC 1998 |
Around 5:00 this afternoon, and again just now, I dialed in to
761-3000 and got "modem answers but no Grex." Each time I re-dialed
761-4931 and got in with no problem. What's going on?
|
other
|
|
response 253 of 306:
|
Jun 9 03:20 UTC 1998 |
Similar. Twice today i dialed in and got the CONNECT 19200 message i always
get, but no welcome and no login prompt. the first time, i tried a couple
times more and then got in. this time i tried a couple times more and then
dialed my isp and telnetted in. on all occasions i dialed 761-3000.
|
jep
|
|
response 254 of 306:
|
Jun 9 13:18 UTC 1998 |
re #251: Nope, it's nice to have someone to talk to. (-:
|
scott
|
|
response 255 of 306:
|
Jun 9 14:24 UTC 1998 |
Grex was crashed last night, not sure when it started.
|
keesan
|
|
response 256 of 306:
|
Jun 9 17:13 UTC 1998 |
I think there is a built-in filter against anyone who cannot figure out how
to use the system, so therefore a shortage of dummies online. I have noticed
how eager people have been to help me, so I must be a real rarity.
|
richard
|
|
response 257 of 306:
|
Jun 12 14:58 UTC 1998 |
I just tried to telnet in and was #73 in the que! (Im still in the que
ten minutes later at #55 as I type this via backtalk) Has grex been
cutting telnet ports or something? Why is there any reason to allow so
many into the que? Wouldnt it be better to cap the que at a max of 35
or 40? Anybody above 35 or 40 in the que, grex could give a:
"sorry, Grex is full at the moment. Try again later or feel free to
visit us via backtalk at www.cyberspace.org"
and then disconnect. I imagine many who find themselves in the 60's or
70's in the que are just ghosts by the time they reach the login prompt.
Any statistics on what percentage reach logon after being in the que but
then dont logon?
|
mcnally
|
|
response 258 of 306:
|
Jun 12 15:42 UTC 1998 |
So basically what you want is a way to get into the queue ahead of others
who you figure probably aren't paying attention by the time their names
come up? If they're gone by the time they get to the head of the queue
the system will take care of it within a minute or two and move on to the
next person.
The whole point of having a queue is so users can get a connection in a
bounded amount of time with connections assigned in an equitable manner --
if the queue "filled up" and you had to compete for a spot in the queue
you could get screwed, with people who started behind you entering the
queue ahead of you -- in fact there'd be little point in *having* the
queue because you'd be right back to the situation where people coming
in would have to "attack dial" (or "attack telnet") for a limited number
of spots and hope to get lucky.
The current system has the advantage of being as fair as it can reasonably
be made and is still quite efficient. I don't think anyone can argue with
the fairness of the queue -- what you're really arguing for is the chance
to have special priveleges and jump ahead of people who have been waiting
longer..
|
rcurl
|
|
response 259 of 306:
|
Jun 12 16:51 UTC 1998 |
It would be useful to know the "telnet attack" frequency after the queue
is full, and also the frequency of timeouts when users reach the head of
the queue, during the course of the day(s). No logical decision on the
optimum length of the queue can be made without knowing its dynamics.
|
davel
|
|
response 260 of 306:
|
Jun 12 17:03 UTC 1998 |
Or, of course, have a queue length of zero, right Richard?
|
jerome
|
|
response 261 of 306:
|
Jun 12 17:27 UTC 1998 |
The queue is definitely a Good Thing, and I think something that Mnet is
sorely lacking. There's been talk of increasing the number of ptys to
allow 80 remote users to log in at once (as opposed to the current setting
of 64). After the mail machine is up, that'll take even more load off the
sun, and could quite possibly permit even more simultaneous users (I'm
looking forward to the day when Grex can support 100 users at once).
I think that increasing the number of simultaenous users allowed (while
being careful to make sure system performance has not been sacraficed)
actually has two positive effects: Not only is the queue going to be
shorter (on the average), but since more users will be able to log in,
check their mail, conferences, etc., and log off, there will also be
a higher turnover of users, and thus the queue will move faster.
|
atticus
|
|
response 262 of 306:
|
Jun 12 18:28 UTC 1998 |
(In fact The Queue is one of the main reasons which persuaded me to
shift to Grex from M-Net.)
|
jep
|
|
response 263 of 306:
|
Jun 12 19:18 UTC 1998 |
I haven't received a notice from M-Net that no ports were available for
many moons.
I also tried to telnet in to Grex earlier, and was shocked to find
myself at a position over 70. The highest I'd seen before was in the
mid-40s. Congrats to Grex on it's success.
However, richard is not entirely off base here. I have telnetted to
Grex and not noticed my position in the queue came up until I had been
disconnected. I don't think many coming here during the day can sit at
their computer and wait until their number comes up, when that can take
15-20 minutes or even much longer. My favorite solution would be to
increase the number of allowed telnet sessions until there are some
complaints about connection speed, but barring that, a limit to length
of the queue doesn't seem out of line to me.
|
remmers
|
|
response 264 of 306:
|
Jun 12 19:34 UTC 1998 |
Unlimited queue length gives greatest flexibility to folks, so I
wouldn't want to see that change. If I connect to Grex and find
myself way back in the queue, I have the choice of either (a)
waiting it out, or (b) disconnecting and trying again later. If
I couldn't even get in the queue, my option would be limited to (b).
How does that help anyone?
When I telnet to Grex, I'm using a windowing environment in which I
can do other things on the interent while I'm waiting in the queue.
Waiting 15-20 minutes or longer is no usually no problem.
I use an ISP which M-Net doesn't always recognize as "local" (i.e.
Michigan-based). The non-local-guest telnet ports on M-Net are
frequently filled up.
|
arianna
|
|
response 265 of 306:
|
Jun 12 20:01 UTC 1998 |
When I see that the queue is obnoxiously high, I just disconnect and wait till
later. *shrug*
|
scg
|
|
response 266 of 306:
|
Jun 12 21:14 UTC 1998 |
If the queue is too long, you can always come back later, which is what would
would have to do without the queue, or with a limit to its length.
M-Net hasn't given me a notice about no ports being available in ages either,
but they consider the network I come in from to be "local", so I have access
to a lot of ports that people coming from other places don't.
|
keesan
|
|
response 267 of 306:
|
Jun 12 21:28 UTC 1998 |
Could people far back in the queue just dial in instead? (If local).
|
scott
|
|
response 268 of 306:
|
Jun 12 22:40 UTC 1998 |
Depends. Some people have ubiquitous Internet (at work or school, usually)
but no modem.
|
keesan
|
|
response 269 of 306:
|
Jun 12 23:38 UTC 1998 |
How does one reach the Internet without a modem?
|
anderyn
|
|
response 270 of 306:
|
Jun 12 23:45 UTC 1998 |
Easy. By turning on my work computer at my desk and clicking on Netscape.
Or by telnetting, or lynxing, etc. It's all connected all the time.
|
scott
|
|
response 271 of 306:
|
Jun 13 01:11 UTC 1998 |
Modems are like bicycles. Businesses and colleges have trucks.
|
keesan
|
|
response 272 of 306:
|
Jun 13 03:52 UTC 1998 |
Jim is trying to explain how there are other types of connections, such as
direct digital to the Internet. I prefer bicycles, less wasteful.
|
mcnally
|
|
response 273 of 306:
|
Jun 13 05:12 UTC 1998 |
Yeah, when I telnet in and get stuck in the queue I just do something in
another window until I hear the beep.. BTW, very good idea putting the
bell character in the login message -- it helps a *lot* when I'm not paying
direct attention. About the only time I miss the login prompt now is when
I step away from the computer or when I'm on a computer that doesn't beep..
|
scg
|
|
response 274 of 306:
|
Jun 13 06:33 UTC 1998 |
I'm not sure I like the truck/bicycle anology, but I guess it fits. For
transporting just you a relatively short distance, your bicycle is far more
efficient than a truck. If you needed to move several hundred pieces of
furniture to another city, the truck would be quite efficient while your
bicycle would take several days to do what the truck could do in an hour.
Imagine an office where you've got 40 people who sometimes need to do stuff
on the Internet. To do that with individual modems, you would need 40 modems,
at around $100 each, for a one time cost of $4,000. You would then need 40
phone lines, at a one time cost of $40 each, for a total of $1,600, and then
around $20 each per month, so $800 per month. Add to that 40 dial-up Internet
access accounts, at around $15 each, for a total of $600 per month. So you're
talking around $5,600 for a startup cost, and then $1,400 per month. Even
that doesn't touch on the personell costs required to support a setup like
that.
Very likely, a single dual channel ISDN connection would be more than adequate
for everything those users are doing. For that sort of connection, which
connects the office ethernet to the Internet, there would be around $800 in
startup costs, including line installation, the router, and the ISP startup
fees, and then around $300 per month for the line and the connection from the
ISP. To top it off, the connection will probably feel a lot faster than the
individual modem connections did.
You're saying the individual modems are less wasteful?
For an example on a smaller scale, there's my home network, which has three
computers on it. One of them has to be on the Internet all the time, because
it's my mail server, so I leave it dialed up over a 33.6 modem link. The
other computers connect to the Internet through that computer, which means
I can run a few network cables in the apartment rather than ordering another
phone line. Again, you're saying that's more wasteful than connecting via
multiple phone lines?
|
