Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!decvax!decwrl!pyramid!nsc!amdahl!dmsd!bass
From: bass@dmsd.UUCP
Newsgroups: net.news.group,net.news.stargate
Subject: Starhub -- one alternative to Stargate
Message-ID: <260@dmsd.UUCP>
Date: Wed, 16-Jul-86 03:56:48 EDT
Article-I.D.: dmsd.260
Posted: Wed Jul 16 03:56:48 1986
Date-Received: Wed, 16-Jul-86 17:31:23 EDT
Distribution: na
Organization: DMS Design, San Luis Obispo Office, CA
Lines: 49
Keywords: Existing technology, cheaper than current 2400 baud compressed
Xref: decwrl net.news.group:5537 net.news.stargate:250
My proposal is simply that we create several hubs with 9600 baud modem service
and ample disk storage to provide store and forward mail service. The modems
will be connected to a buffer box with 1 to 2mb of ram storage to match
communications throughput limits at each client machine. The buffer box will
normally be transparent between the modem and system allowing normal modem
service. During preset times when the modem is idle, the buffer box will
call one of the Starhubs and stream in a news batch at full line speed, AND
login to the client system and dump the stream to the unbatcher matching client
speeds.
Current DDS AT&T nite service for such a news feed would be about $75/month
and hub service fees of less than $50/month. The buffer box will cost between
$300 and $500, the modem between $1,000 and $1,500 depending on what can be
done for a group buy. The DDS rate can be reduced maybe 50% with a group
or volume discount.
The Hub could provide inwatts service and bill-back all the charges in
one bill. A single $9,000 hub could support about 80 systems each night
and another 100 systems during the day/evening. Hub based service
could be both faster than the current backbone, more reliable, and cheaper
over all -- and remove a major bias in the cost of usenet caused by the
backbone. The total cost of usenet would be reduced as much as 70% from the
current implementation. Hubs operated as a fee based service would pay back
in about 12 months. Between 6 and 10 hubs could service most of the major
cities and surrounding area.
OR -- the hubs need not exist at all .... using the same technology with
the current backbone. Due to the higher connectivity of the current backbone
over a multiple hub arrangment, the costs will be higher than hubs,
be still significantly cheaper than current 1200/2400 DDS service.
Current DDS AT&T nite service for news costs about $360/month assuming a
2400 baud compressed news feed (plus/minus about $60). The capital investment
is a 2400 baud modem costing less than $1,000. Many sites are using 1200 baud
compressed service at twice the cost (why I don't know, since the capital
payback is about 2 months).
Normal uucp is limited to below 4800 baud average throughput due to system
timings on many machines, which limits the direct application of 9600/19.2kb
modems for uucp based service. But even still, at 4800 baud they will still
pay back over less than 6 months.
The software to implement this is about 6 man months, plus 3 man months
for the buffer box firmware and hardware.
--
John Bass (DBA:DMS Design)
DMS Design (System Design, Performance and Arch Consultants)
{dual,fortune,polyslo,hpda}!dmsd!bass (805) 541-1575
Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!henry
From: henry@utzoo.UUCP (Henry Spencer)
Newsgroups: net.news.group,net.news.stargate
Subject: Re: Starhub -- one alternative to Stargate
Message-ID: <6945@utzoo.UUCP>
Date: Wed, 16-Jul-86 18:34:30 EDT
Article-I.D.: utzoo.6945
Posted: Wed Jul 16 18:34:30 1986
Date-Received: Wed, 16-Jul-86 18:34:30 EDT
References: <260@dmsd.UUCP>
Organization: U of Toronto Zoology
Lines: 15
> ...The capital investment
> is a 2400 baud modem costing less than $1,000. Many sites are using 1200 baud
> compressed service at twice the cost (why I don't know, since the capital
> payback is about 2 months).
It's really easy to figure out, actually: phone-bill money and capital-
equipment money are not freely interconvertible in most large organizations.
Oh, they are if you go high enough up the management tree... but quite often
Usenet cannot stand the sort of upper-management scrutiny that would result.
There is also a chicken-and-egg problem, since *your* investment does not
pay back a cent until your neighbors make similar investments.
--
Henry Spencer @ U of Toronto Zoology
{allegra,ihnp4,decvax,pyramid}!utzoo!henry
Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!caip!sri-spam!nike!cit-vax!amdahl!dmsd!bass
From: bass@dmsd.UUCP (John Bass)
Newsgroups: net.news.group,net.news.stargate
Subject: Re: Starhub -- one alternative to Stargate
Message-ID: <262@dmsd.UUCP>
Date: Fri, 18-Jul-86 13:00:26 EDT
Article-I.D.: dmsd.262
Posted: Fri Jul 18 13:00:26 1986
Date-Received: Sun, 20-Jul-86 04:39:10 EDT
References: <260@dmsd.UUCP> <6945@utzoo.UUCP>
Organization: DMS Design, San Luis Obispo Office, CA
Lines: 20
Xref: watmath net.news.group:6183 net.news.stargate:268
Summary: One the otherhand .....
In article <6945@utzoo.UUCP>, henry@utzoo.UUCP (Henry Spencer) writes:
> It's really easy to figure out, actually: phone-bill money and capital-
> equipment money are not freely interconvertible in most large organizations.
> Oh, they are if you go high enough up the management tree... but quite often
> Usenet cannot stand the sort of upper-management scrutiny that would result.
Interesting you should note this, the same problem only worse is true with
Stargate .... purchase/rental of buffer box, converter, plus installation
charges for cable, cable service fees, Stargate service fees, and so on ...
Atlease with 9600 baud modem technology it has a dual use of increasing
productivity with home/remote installations and all the other uses
modems have ....
Which is easier to justify????
--
John Bass (DBA:DMS Design)
DMS Design (System Design, Performance and Arch Consultants)
{dual,fortune,polyslo,hpda}!dmsd!bass (805) 541-1575
Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!henry
From: henry@utzoo.UUCP (Henry Spencer)
Newsgroups: net.news.group,net.news.stargate
Subject: Re: Starhub -- one alternative to Stargate
Message-ID: <6970@utzoo.UUCP>
Date: Mon, 21-Jul-86 17:00:18 EDT
Article-I.D.: utzoo.6970
Posted: Mon Jul 21 17:00:18 1986
Date-Received: Mon, 21-Jul-86 17:00:18 EDT
References: <260@dmsd.UUCP> <6945@utzoo.UUCP>, <262@dmsd.UUCP>
Organization: U of Toronto Zoology
Lines: 31
> > It's really easy to figure out, actually: phone-bill money and capital-
> > equipment money are not freely interconvertible in most large organizations.
>
> Interesting you should note this, the same problem only worse is true with
> Stargate ...
Quite true. It will be true of *any* scheme that offers improved service
at the cost of buying equipment. But on a moderated channel like Stargate,
we won't have to do it over again two years later when our 9600-baud modems
are saturated and we need to upgrade to 19.2.
> Atlease with 9600 baud modem technology it has a dual use of increasing
> productivity with home/remote installations and all the other uses
> modems have ....
>
> Which is easier to justify????
Neither. How many home/remote installations do you know of that have 9600
baud modems and want to talk to you that way? None, right. Although this
isn't something I'm sure of, I would guess that the vast majority of the
2400-baud modems on the net were bought to help keep news under control,
not because the dialup community was clamoring for them.
Actually, I agree that "dual use" is a useful *excuse*. It's just that it
wears thin when used for the fifth time in five years, with no end in sight.
We need to control the volume, not constantly struggle to accommodate more
and more and more.
--
EDEC: Stupidly non-standard
brain-damaged incompatible Henry Spencer @ U of Toronto Zoology
proprietary protocol used. {allegra,ihnp4,decvax,pyramid}!utzoo!henry
Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!henry
From: henry@utzoo.UUCP (Henry Spencer)
Newsgroups: net.news.group,net.news.stargate
Subject: Re: Starhub -- one alternative to Stargate
Message-ID: <6971@utzoo.UUCP>
Date: Mon, 21-Jul-86 17:04:01 EDT
Article-I.D.: utzoo.6971
Posted: Mon Jul 21 17:04:01 1986
Date-Received: Mon, 21-Jul-86 17:04:01 EDT
References: <260@dmsd.UUCP> <6945@utzoo.UUCP>, <262@dmsd.UUCP>
Organization: U of Toronto Zoology
Lines: 10
A further question about Starhub: how do you plan to address the legal
issues of liability for content? Any centralized organization, be it
Stargate or Starhub, is a potential target for lawsuits. Stargate, last
I heard, plans to address this basically by being an all-moderated net.
An organization which plans to pass all traffic that comes in is going
to be paying an awful lot for liability insurance.
--
EDEC: Stupidly non-standard
brain-damaged incompatible Henry Spencer @ U of Toronto Zoology
proprietary protocol used. {allegra,ihnp4,decvax,pyramid}!utzoo!henry
Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!lll-crg!hoptoad!gnu
From: gnu@hoptoad.uucp (John Gilmore)
Newsgroups: net.news.group,net.news.stargate
Subject: What makes Stargate moderation work better?
Message-ID: <917@hoptoad.uucp>
Date: Fri, 25-Jul-86 11:26:27 EDT
Article-I.D.: hoptoad.917
Posted: Fri Jul 25 11:26:27 1986
Date-Received: Fri, 25-Jul-86 21:34:08 EDT
References: <260@dmsd.UUCP> <6945@utzoo.UUCP>, <262@dmsd.UUCP> <6970@utzoo.UUCP>
Organization: Nebula Consultants in San Francisco
Lines: 30
Xref: mnetor net.news.group:3272 net.news.stargate:208
In article <6970@utzoo.UUCP>, henry@utzoo.UUCP (Henry Spencer) writes:
> ...on a moderated channel like Stargate,
> we won't have to do it over again two years later when our 9600-baud modems
> are saturated and we need to upgrade to 19.2.
Gee, that doesn't take a Stargate or a Starhub or anything.
Just remove "net,world,comp,..." from your sys file and leave in "mod".
Boy that was simple and it doesn't even need next year's technology!
Oh, I see...stargate will somehow moderate things that the current
Usenet would not? With paid staff or what? We seem to have all the
volunteer moderators busy already; at least I don't see anybody
screaming for the job of mod.unix-wizards or mod.lang.c moderator.
(Let alone the flame groups.)
These days I am not so sure that moderation (editorial functions) would
reduce the traffic significantly. Use a big figure and say that 50%
of the current traffic is trash that an editor would weed out. Now we're
back to last year's traffic level. We've bought ourselves a year. Why won't
the traffic grow back to today's level again as more machines and users
and gateways and interesting topics join the net? Or are the moderators
going to disallow new topics, new gateways, new users?
I think Usenet traffic is like files on disk: it grows to fill all the
available space and then you have to prune periodically. But there's
no way that today's actions will keep it from filling up next month or next
year.
--
John Gilmore {sun,ptsfa,lll-crg,ihnp4}!hoptoad!gnu jgil...@lll-crg.arpa
May the Source be with you!
Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!henry
From: henry@utzoo.UUCP (Henry Spencer)
Newsgroups: net.news.group,net.news.stargate
Subject: Re: What makes Stargate moderation work better?
Message-ID: <7011@utzoo.UUCP>
Date: Thu, 31-Jul-86 17:18:37 EDT
Article-I.D.: utzoo.7011
Posted: Thu Jul 31 17:18:37 1986
Date-Received: Thu, 31-Jul-86 17:18:37 EDT
References: <260@dmsd.UUCP> <6945@utzoo.UUCP>, <262@dmsd.UUCP>
<6970@utzoo.UUCP>, <917@hoptoad.uucp>
Organization: U of Toronto Zoology
Lines: 70
> Gee, that doesn't take a Stargate or a Starhub or anything.
> Just remove "net,world,comp,..." from your sys file and leave in "mod".
> Boy that was simple and it doesn't even need next year's technology!
Unfortunately, it cuts one off from the rest of the community to a large
extent unless everybody does the same thing at the same time. The chances
of getting unanimity on this are nil. The result is a considerable
penalty for doing it first, which is why there is a slight shortage of
people volunteering to do it. Stargate's virtue here is that everyone
starts out equal.
> Oh, I see...stargate will somehow moderate things that the current
> Usenet would not?
Yes, easily. The current Usenet has a major chicken-and-egg problem:
*most* of the groups could be and should be moderated, but as witness
mod.unix, it's often hard to get a moderated group started when there
is an unmoderated alternative. Much of the readership would probably
prefer a moderated version of (say) net.lang.c, but nobody will post
to it if all the readers are still over in the unmoderated group.
> With paid staff or what? We seem to have all the
> volunteer moderators busy already; at least I don't see anybody
> screaming for the job of mod.unix-wizards or mod.lang.c moderator.
My understanding is that Stargate quite definitely intends to pay its
moderators. It's too much work for volunteer labor, which is the reason
why we're a little short of volunteer moderators.
> These days I am not so sure that moderation (editorial functions) would
> reduce the traffic significantly. Use a big figure and say that 50%
> of the current traffic is trash that an editor would weed out...
Personally, I'd put the percentage a lot higher than that. Furthermore,
the percentage is growing. The signal/noise ratio is visibly dropping
in most of the unmoderated groups.
> Now we're back to
> last year's traffic level. We've bought ourselves a year. Why won't
> the traffic grow back to today's level again as more machines and users
> and gateways and interesting topics join the net? Or are the moderators
> going to disallow new topics, new gateways, new users?
Even stipulating for a moment that the signal/noise ratio remains steady,
there are *three* things you can do when the pool of contributions grows:
(a) Leave the rejection threshold where it is and let the volume grow.
(b) Exclude new material and new people, holding volume constant by
simply stifling growth.
(c) Raise the rejection threshold, holding volume constant by demanding
higher quality.
Choice (a) is roughly what we've got now, with threshold at zero. Choice
(b) isn't an answer at all. Choice (c) is clearly the one most readers
would like, if for no other reason than that it gives them better value
for the time they spend reading.
In a network of finite bandwidth, sooner or later there has to come a
limit. If traffic is exceeding the limit, the moderators have to raise
their threshold of what's interesting to the readership. This does *not*
mean disallowing new topics, new gateways, or new users; it means that
even the old contributors, using old gateways to discuss old topics,
will get their contributions bounced when they are boring, irrelevant,
repetitive, or stupid. Sounds good to me.
--
EDEC: Stupidly non-standard
brain-damaged incompatible Henry Spencer @ U of Toronto Zoology
proprietary protocol used. {allegra,ihnp4,decvax,pyramid}!utzoo!henry
|