|
Grex > Coop11 > #32: How should we determine how many dialin lines we should have? | |
|
| Author |
Message |
| 25 new of 154 responses total. |
scg
|
|
response 75 of 154:
| 
May 21 04:34 UTC 2000 |
The Net connections aren't a scarce resource. The dial-ups are.
|
aruba
|
|
response 76 of 154:
|
May 21 15:09 UTC 2000 |
At the moment, that's not true, Steve. I dialup most of the time too.
|
scg
|
|
response 77 of 154:
|
May 21 15:41 UTC 2000 |
Ok, let me rephrase that. The Net connections aren't an expensive resource.
The dial-ups are.
The limit on the number of incoming telnet connections is an entirely
artificial limit. It was a matter of taking how many simultanious users
somebody felt Grex could support well, and subtracting the number of dial-up
lines. Raising that is a matter of changing a configuration file.
The dial-up lines cost at least $20 each. Any dial-up line we can get rid
of means the formula that was used to set the number of incoming telnet users
would call for increasing the number of telnet users. For that matter, lower
usage of the dial-ups, even if they're still there, would call for increasing
the number of telnet users.
|
cmcgee
|
|
response 78 of 154:
|
May 21 16:07 UTC 2000 |
Some of us use dialups so we have internet access from our home. Otherwise
its a trip to the library and a time-limited access there, only during
business hours (which, for example, will not include Sunday starting next
week).
|
scg
|
|
response 79 of 154:
|
May 21 17:10 UTC 2000 |
Right. I'm not saying that the dial-ups are useless, just that they're more
expensive to provide. Clearly we get some benefit out of having some number
of dial-up lines.
|
gull
|
|
response 80 of 154:
|
May 22 05:58 UTC 2000 |
So basically, people who dial in because they have to are fine. People who
could telnet in, but dial in instead to avoid the queue, are effectively
costing Grex money. (Not directly, of course, but effectively by making it
look like we need more dialin lines than we really do.)
|
other
|
|
response 81 of 154:
|
May 22 06:17 UTC 2000 |
That is not an accurate representation of the facts, though it is a
convincingly damning interpretation.
|
scg
|
|
response 82 of 154:
|
May 22 06:22 UTC 2000 |
I think the way I would explain it is this: Everybody who dials in costs
Grex more money to support them than somebody who telnets in. That's a fact.
Given that fact, whether the users we gain by having the dial-ups are worth
more to us than the cost of the dial-ups is a question. I believe the answer
to that question is yes.
|
albaugh
|
|
response 83 of 154:
|
May 22 21:58 UTC 2000 |
What "fact"?! Does grex pay ma bell by the connection minute? I think not!
Nobody is costing grex anything more than grex chooses to provide. It's a
resource. It costs the same whether anyone uses a dialup or not. So just
get off the whole "users costing grex more" argument.
|
scg
|
|
response 84 of 154:
|
May 22 22:07 UTC 2000 |
If Grex bases its number of dial-in lines on the number needed to support the
dial-in user base, than Grex does pay Ameritech based on how much time users
are spending dialed in on them. If Grex were instead to buy a several hundred
line huntgroup without consideration for its use you might have a point, but
we're not, and you don't.
I still don't understand why a few people get so upset when I point out that
it costs more to suport dial-up users than to support users coming over the
Net connection. I'm certainly not arguing that the expense isn't worth it.
|
albaugh
|
|
response 85 of 154:
|
May 22 22:09 UTC 2000 |
Just don't use phrasing that "people who dial in are costing grex more".
People don't cost grex a damn thing. Grex deterines the resources, people
use 'em.
|
scg
|
|
response 86 of 154:
|
May 22 23:05 UTC 2000 |
Are you offended by my phrasing?
Grex hopefully determines the resources based at least in part on whether they
are going to be used.
|
devnull
|
|
response 87 of 154:
|
May 23 01:32 UTC 2000 |
Re #80: Exactly, and it seems to me that if there are significant numbers
of such people, we ought to find some way for those people to be able to
telnet in with the same level of priority that they get when they dial in.
I'm also not sure I'm convinced that going for as few busy signals as
grex gets is a particularily appropriate goal, but I can see both sides
of the argument, and I'm not convinced that the status quo is horible.
|
mdw
|
|
response 88 of 154:
|
May 23 03:30 UTC 2000 |
Re #82, Steve forgot to mention that the same user population that is
more likely to use dial-ups is also more likely to contribute (become a
member) of grex. However, as internet access is becoming more
universal, this isn't as true as it once was.
|
davel
|
|
response 89 of 154:
|
May 23 11:21 UTC 2000 |
I rarely dial in to avoid the queue. From home, I have no ISP, and a dialup
is the only real way to access Grex.
|
gelinas
|
|
response 90 of 154:
|
May 23 18:48 UTC 2000 |
Which is why we need to continue with modems. The question is, how *many*
modems do we need? Can we cut back on the number and still provide the same
level of service?
From what I've seen so far, we can.
|
hhsrat
|
|
response 91 of 154:
|
May 24 20:24 UTC 2000 |
Another point: The less dial-up numbers we have, the less modems we need
to use. The less modems we use, the less power we use. Compared to our
overall power usage, modem-drawn power may be negligible, but if the
modems are power-hungry, getting rid of even 1 dial up line COULD result
in savings on our electric bill.
|
scott
|
|
response 92 of 154:
|
May 24 21:12 UTC 2000 |
Very small savings from electricity.
|
keesan
|
|
response 93 of 154:
|
May 29 23:05 UTC 2000 |
I can be on grex by dialing within 30 seconds of turning on my computer. If
I were to use some browser to access an ISP, and then telnet, it would take
me several minutes longer to get onto grex, and then I get the impression that
things just take longer - I type a word and it does not appear on the screen
as quickly, having to go through the net. I have telnetted to grex from
Netscape 3 and 4 and from Arachne, but I see no advantage at all to direct
dialing with Procomm (which gives me a nice legible screen, too).
|
gull
|
|
response 94 of 154:
|
May 30 03:21 UTC 2000 |
Yes, telnet connections have much higher latency, especially given the
overworked charater of Grex's internet link. I'm usually typing three to
four words ahead of the display when I'm entering a response.
|
mooncat
|
|
response 95 of 154:
|
May 30 15:21 UTC 2000 |
I could dial-i, but I chose not to most of the time. Mainly because I
talk to people via IMs on AOL and messaging back and forth on ICQ. By
telneting in I can be talking to multiple people in multiple places. I
generally don't run across too much trouble with lag, just every once
and awhile AOL gets weird and boots me- but that's obviously not a grex
problem. <grins>
I still get the same legible screens (although for some reason I have a
problem reading conferences when I'm telnetted in) and see several
benefits to being telnetted in.
|
mwg
|
|
response 96 of 154:
|
Jun 6 16:50 UTC 2000 |
I telnet in from home, and legibility is a non-issue because I am running
Linux and telnet from a text-mode console. My biggest legibility issues
from being at the office, where I have to use Windows.
Dialing in is not a serious option for me as I live in Southgate, and A2
is a bit of a haul on a land-line. As for the net.wierdness, it has
always been there, so I don't notice any more, unless it stops completely
or gets VERY delayed.
|
janc
|
|
response 97 of 154:
|
Jul 23 05:20 UTC 2000 |
Some new statistics. I think we still have 11 dial-in lines.
Usage between Mon May 1 00:00:00 2000 and Wed May 31 23:59:59 2000
of IP addresses: 204.212.46.132
0: 220.10 29.59%
1: 248.34 33.38%
2: 154.62 20.78%
3: 75.20 10.11%
4: 32.55 4.38%
5: 10.37 1.39%
6: 1.87 0.25%
7: 0.79 0.11%
8: 0.09 0.01%
9: 0 0
10: 0 0
11: 0 0
Usage between Thu Jun 1 00:00:00 2000 and Fri Jun 30 23:59:59 2000
of IP addresses: 204.212.46.132
0: 162.11 22.52%
1: 199.08 27.65%
2: 155.13 21.55%
3: 108.48 15.07%
4: 54.13 7.52%
5: 25.83 3.59%
6: 9.42 1.31%
7: 3.99 0.55%
8: 0.92 0.13%
9: 0.55 0.08%
10: 0.29 0.04%
11: 0 0
Usage between Sat Jul 1 00:00:00 2000 and Sun Jul 23 01:19:07 EDT 2000
of IP addresses: 204.212.46.132
0: 163.26 30.87%
1: 175.01 33.09%
2: 109.09 20.63%
3: 52.99 10.02%
4: 18.09 3.42%
5: 6.98 1.32%
6: 2.20 0.42%
7: 0.92 0.17%
8: 0.29 0.05%
9: 0.02 0.00%
10: 0 0
11: 0 0
Dial-in use was up a smidgeon in June, but I think that was because of the
M-Net outage.
I think we could be very happy with 6 lines.
|
prp
|
|
response 98 of 154:
|
Jul 23 21:13 UTC 2000 |
And at about $20/mo five lines is $100, or about enough to double the
bandwidth of the Internet connection. That should be good enough to
eliminate the login queue.
|
i
|
|
response 99 of 154:
|
Jul 24 02:19 UTC 2000 |
Isn't the login queue due to a limited supply of virtual tty's on grex
to log in to? Logging in through the internet from work of late, i have
found the keystroke delay rather bad at times, and boosting our internet
bandwidth could help on that front.
|