mra@searchtech.com (Michael Almond) (03/11/91)
PSI already mentioned that they are placing the leaf node restrictions only on the $75/month UUPSI customers. They, PSI, offer a second grade of service, with a higher price, that allows unrestricted UUPSI access. Why should this bother anyone? It sounds to me like PSI found that $75 couldn't cover the cost of non-leaf connections and needed to change their service a little. Boycott? Sounds a little extreme for a rate change. -- Michael R. Almond (Georgia Tech Alumnus) mra@srchtec.uucp (registered) search technology, inc. mra@searchtech.com 4725 peachtree corners cir., suite 200 {uupsi,stiatl}!srchtec!mra norcross, georgia 30092 (404) 441-1457 (office)
murray@sun13.scri.fsu.edu (John Murray) (03/12/91)
In article <1991Mar11.143824.24170@searchtech.com> mra@searchtech.com (Michael Almond) writes: >PSI already mentioned that they are placing the leaf node restrictions only >on the $75/month UUPSI customers. They, PSI, offer a second grade of service, >with a higher price, that allows unrestricted UUPSI access. > >Why should this bother anyone? It sounds to me like PSI found that $75 >couldn't cover the cost of non-leaf connections and needed to change their >service a little. Did you miss the post that mentioned how much this other service cost? First, think about this: In a newsfeed, what's more costly, providing the downlink-side (the rest of the world) of a full or sizeable feed to a leaf site or (relatively) small cluster of sites, or providing the uplink side for all the posts that originate from that site or sites? Well, lessee, 15 MB/day for the down side, and an average of (15 MB / 18000 sites <est.> ) * number of sites in this cluster, for the up side. If this little cluster is, say 100 sites (!) the uplink side is still well under 1% of the total cost in bandwidth. Mail volume ignored, since even a high-volume mailing list or two run from this cluster isn't going to put a dent in these numbers. Cost for leaf site: $75. Cost for full site: exstimated at a little under $300. Liability for stuff posted by non-customers is a NON-ISSUE, since it would be TRIVIAL for PSI to contractually put the burden of liability on the customer that fed them the "bad stuff". (of course, we might jump on them for that, too };-> ) I had no opinion on this issue until one person mentioned what the cost for an unrestricted site was. (BTW, can someone verify the $825-875 quarterly cost for a full site?) Now I believe that this move is strictly a *marketing* based decision on the part of PSI. >Michael R. Almond (Georgia Tech Alumnus) mra@srchtec.uucp (registered) -- Disclaimer: Yeah, right, like you really believe I run this place. John R. Murray | "Memory serves murray@vsjrm.scri.fsu.edu | wise commanders." Supercomputer Research Inst.| - Tz'u-hsi, 638 AD
xanthian@zorch.SF-Bay.ORG (Kent Paul Dolan) (03/12/91)
murray@sun13.scri.fsu.edu (John Murray) writes: > mra@searchtech.com (Michael Almond) writes: >> PSI already mentioned that they are placing the leaf node >> restrictions only on the $75/month UUPSI customers. They, PSI, offer >> a second grade of service, with a higher price, that allows >> unrestricted UUPSI access. >> Why should this bother anyone? It sounds to me like PSI found that >> $75 couldn't cover the cost of non-leaf connections and needed to >> change their service a little. > Did you miss the post that mentioned how much this other service cost? > First, think about this: In a newsfeed, what's more costly, providing > the downlink-side (the rest of the world) of a full or sizeable feed > to a leaf site or (relatively) small cluster of sites, or providing > the uplink side for all the posts that originate from that site or > sites? Well, lessee, 15 MB/day for the down side, and an average of > (15 MB / 18000 sites <est.> ) * number of sites in this cluster, for > the up side. If this little cluster is, say 100 sites (!) the uplink > side is still well under 1% of the total cost in bandwidth. Mail > volume ignored, since even a high-volume mailing list or two run from > this cluster isn't going to put a dent in these numbers. > Cost for leaf site: $75. Cost for full site: exstimated at a little > under $300. > I had no opinion on this issue until one person mentioned what the > cost for an unrestricted site was. (BTW, can someone verify the > $825-875 quarterly cost for a full site?) Now I believe that this move > is strictly a *marketing* based decision on the part of PSI. Absolutely correct; they are trying to _force_ each customer that wants to take advantage of this cheap feed to connect to them directly and pay full freight (and fight over bandwidth, while UUPSI build an empire trying to keep up, the same problem UUNET is displaying today, and absolutely typical of reacting to problems rather than forecasting them and being ready before they can occur). If the problem were really upload traffic, they could simply limit uploads to 5% of downloads (a ridiculously generous allowance), which would make sure they weren't being hurt transporting an unfair amount of garbage compared to what the customer was buying in terms of bandwidth. Instead of this reasonable approach, they have chosen to dictate to their customers the type of connectivity those customers will display across the customer-UUPSI interface, which frankly is none of their damned business. Since they are selling connectivity rather than bandwidth, what their product really amounts to for the not-terribly-clever-about-news-software customer is a perpetual bondage into leafhood. Trying to convince those who have watched the net grow that it is going to behave the same with a large class of second class citizens who can't extend a feed to a friend in the time honored manner of the net without quadrupling the cost of a feed is a waste of time by all those attempting to defend UUPSI in this matter. This is bad juju for the net, period. Kent, the man from xanth. <xanthian@Zorch.SF-Bay.ORG> <xanthian@well.sf.ca.us>
mpd@anomaly.SBS.COM (Michael P. Deignan) (03/13/91)
murray@sun13.scri.fsu.edu (John Murray) writes: >First, think about this: In a newsfeed, what's more costly, providing the >downlink-side (the rest of the world) of a full or sizeable feed to a >leaf site or (relatively) small cluster of sites, or providing the >uplink side for all the posts that originate from that site or sites? >Well, lessee, 15 MB/day for the down side, and an average of >(15 MB / 18000 sites <est.> ) * number of sites in this cluster, for the >up side. If this little cluster is, say 100 sites (!) the uplink side is >still well under 1% of the total cost in bandwidth. Mail volume ignored, >since even a high-volume mailing list or two run from this cluster isn't >going to put a dent in these numbers. Praise Gene Spafford! Someone finally posted a message about volume! Clearly, the volume inbound to PSI from their "leaf sites" could, at most, be 15 megabytes, if the maximum outbound volume was 15 megabytes. So, it would seem to lead one to the conclusion that at most a completely unrestricted inbound/outbound feed should cost no more than $150. ($75 per inbound/outbound side) However, most sites do not feed PSI 15megs of inbound traffic. Otherwise, PSI wouldn't have 15mb, they would have 15mb * # of sites they feed. So, the amount of inbound traffic per site is clearly minimal, or roughly the 15mb / # sites they feed - which is really an overstatement, since PSI may get their news from non-subscribers like UUNET, where PSI pays for the news. Of course, with some subscribers, this might not be entirely true, since a university would have a potentially higher mail volume than a "normal" site. Then again, how many universities use PSI's $75/mo connection? What we have here is an attempt to limit the amount of inbound traffic by PIS in an attempt to make unconnected sites pay $75 for their own mail/news inbound feed. (Why doesn't PSI offer an "inbound only" service for, say, $25/mo?) Why, that would cut into the "bottom line", by gosh! In PSI offered a complete "non-restrictive" news/mail feed, for slightly more money (like, $100/mo) to cover the inbound "non-leaf-site" mail/news, this would probably be acceptable to most customers. MD -- -- Michael P. Deignan / -- Domain: mpd@anomaly.sbs.com / "The Mother Of All Battles" -- UUCP: ...!uunet!rayssd!anomaly!mpd / apparently had an abortion... -- Telebit: +1 401 455 0347 /
stanley@phoenix.com (John Stanley) (03/13/91)
mra@searchtech.com (Michael Almond) writes: > PSI already mentioned that they are placing the leaf node restrictions only > on the $75/month UUPSI customers. They, PSI, offer a second grade of service > with a higher price, that allows unrestricted UUPSI access. The next level of service from PSI is $175/month, and allows dialup Internet (SLIP) access. This is more than twice the cost of the USENET News service. The $75/month service was promoted as giving unrestricted uupsi access. > Why should this bother anyone? It sounds to me like PSI found that $75 > couldn't cover the cost of non-leaf connections and needed to change their > service a little. When I signed up for PSI service, they (and I) went to the effort to obtain a domain name for this system. What purpose a domain name? It makes mail from third parties easier to send. Instead of knowing a UUCP routing, they tack on the domain name and Bingo! the mail shows up here. Now PSI is saying (to some) that you may neither send nor receive mail through PSI involving third parties. Now what use is a domain name? If PSI did not intend to allow third party mail to be sent or recieved through uupsi, why was getting a domain name one of the features of their service? Why do they MX, if all they will allow is mail from PSI? As far as news goes, I think it has already been pointed out that they will either see it when I feed it upstream, or they will see it when they feed it back downstream to me. When they feed it downstream to me, it will be larger (more elements in the Path:), and older, which makes their service look slower. There has also been a description of how to get around any automatic news chopping (which will not work at this site). > Boycott? Sounds a little extreme for a rate change. Boycott sounds about right for bait and switch. A user signs up for the USENET feed. When he decides to carry on the USENET tradition of providing feeds to others, he is told he needs to buy a different (SLIP) feed, for more money. I would suggest to PSI that they examine this decision to enforce leaf status again. They will be carrying the same news in either case. I cannot believe that the amount of mail generated by any reasonable node I feed will make any dent in their capacity. In fact, any node I feed that starts to make a dent in MY capacity will be advised to talk to UUNET (would have been PSI, until this came up). On the other hand, the amount of negative publicity they have gotten (in the very newsgroups they want to sell) cannot be beneficial to them. If PSI is really concerned about the load on their systems, perhaps they would be best served by staunching the flow of unrequested newsgroups. The junk newsgroup is the most active one, here. I have not yet received the letter from PSI, so all I am going on is what has been posted here. I am moderately unhappy with the news limitation, but can live with it. However, the first piece of mail from a third party that they drop on the floor will be announced worldwide. They had best not make the mistake of dropping mail I send to myself from an outside system, and it will be impossible to tell this mail from true third party mail.
tneff@bfmny0.BFM.COM (Tom Neff) (03/13/91)
In article <Dccqy4w163w@phoenix.com> stanley@phoenix.com (John Stanley) writes: > Now PSI is saying (to some) that you may neither send nor receive >mail through PSI involving third parties. Now what use is a domain name? >If PSI did not intend to allow third party mail to be sent or recieved >through uupsi, why was getting a domain name one of the features of >their service? Why do they MX, if all they will allow is mail from PSI? Hehe. That is indeed what the stupidly worded contract says, isn't it. Of course if you can't send mail TO or FROM anyone else but UUPSI, then your "mail feed" is worthless, so we presume that's not what PSI "really meant," despite what their contract _says_. Therefore PSI's customers are in the wonderful bind of having to violate the letter of their contract every day, relying instead on PSI's voluntary goodwill in enforcing an informal, verbal reinterpretation of the signed contract. If you want to run a business relationship on "never mind what I say, just do what I tell you," why bother with messy old contracts to begin with? Ain't chaos wonderful. As has been suggested several times, it would be a piece of cake to circumvent PSI's restrictions, but... to quote a sainted former President... <hunch shoulders, waggle V-signs> "That would be wrong!"
peter@sugar.hackercorp.com (Peter da Silva) (03/13/91)
In article <1991Mar12.134431.14294@zorch.SF-Bay.ORG> xanthian@zorch.SF-Bay.ORG (Kent Paul Dolan) writes: > Absolutely correct; they are trying to _force_ each customer that wants > to take advantage of this cheap feed to connect to them directly and > pay full freight Absolutely correct. So? They priced the cheap feed with the expectation that a reasonable percentage of the sites in an area would pay the $75 for a direct feed to UUPSI, to pay for the cost of the dedicated links. When it became obvious that that wasn't going to happen, they changed the rules to further encourage a direct UUPSI connect. If they don't do something like this, the limiting case comes when each area and its expensive dedicated line serves a single UUPSI customer who then feeds all the other sites. This is a perennial problem with Usenet, where whever can ship the most bits the fastest becomes "the" feed for an area, the redundant feeds drop off, and when "the" feed goes everyone panics. It's happened twice in the Houston-Dallas-Austin area that I can remember. And people still don't seem to be learning. > Instead of this reasonable approach, they have chosen to dictate to > their customers the type of connectivity those customers will display > across the customer-UUPSI interface, which frankly is none of their > damned business. No, they're dictating whose traffic will go over the link. It's no different from (and, if anything, more liberal than) any commercial network. I get a flat rate from SWBell but, common carrier status notwitstanding, I'm sure they'd drop me in a minute if I tried running wires into my neighbors' apartments and set up some sort of party-line mechanism. > Since they are selling connectivity rather than bandwidth, what their > product really amounts to for the not-terribly-clever-about-news-software > customer is a perpetual bondage into leafhood. No it *doesn't*. There is absolutely no reason a UUPSI customer has to be a leaf. I *know* you know better than that, Kent. I have explained several times in Email and on the net how to set up a UUPSI feed so you can have a well connected site and still take advantage of their service. Anyone with the brains to set up C News should be able to do it. As you pointed out, the reverse flow is *cheap*. Surely one site in an area can afford the long-distance calls to get news and mail back to any of dozens of well-connected sites that are happy to take them. If not, you can all get together and order a UUNET connect. Or depend on the existing usenet flood algorithm to get the message back. That's what you seem to be pushing for: for UUPSI to get out of the pool and let us go back to the uncertanties of riding in the backs of big sites. > This is bad juju for the net, period. No, Kent. Helping people get off the tax-subsidised backs of universities and the government, and quit hanging on the uncertain arms of friendly system administrators in a hostile commercial environment, is good juju. Period. -- Peter da Silva. `-_-' <peter@sugar.hackercorp.com>.
xanthian@zorch.SF-Bay.ORG (Kent Paul Dolan) (03/14/91)
peter@sugar.hackercorp.com (Peter da Silva) writes: > xanthian@zorch.SF-Bay.ORG (Kent Paul Dolan) writes: >> Since they are selling connectivity rather than bandwidth, what their >> product really amounts to for the not-terribly-clever-about-news-software ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ >> customer is a perpetual bondage into leafhood. > No it *doesn't*. There is absolutely no reason a UUPSI customer has to > be a leaf. I *know* you know better than that, Kent. I have explained > several times in Email and on the net how to set up a UUPSI feed so > you can have a well connected site and still take advantage of their > service. Anyone with the brains to set up C News should be able to do > it. Peter, if you would ever LEARN TO READ THE ENGLISH LANGUAGE, you could avoid dozens of flamefests into which you plunge youself each month by bulling ahead on your own path, blind eyed, brain in neutral, without bothering to read what you are answering. You don't put yourself in the right by saying the same stupid, wrongheaded, ignorant thing more times in a row than anyone is willing to answer, no matter how convinced you are to the contrary, no matter how often you act as if that will work, no matter how many people start nodding their heads hypnotized after their senses go numb from your continued abuse. Posting the same idiocy that gets rejected in email just proves you to be an exceptionally, perhaps uniquely, slow learner. Kent, the man from xanth. <xanthian@Zorch.SF-Bay.ORG> <xanthian@well.sf.ca.us>
john@jwt.UUCP (John Temples) (03/14/91)
In article <1991Mar13.012437.24023@anomaly.SBS.COM> mpd@anomaly.SBS.COM (Michael P. Deignan) writes: >Why doesn't PSI offer an "inbound only" service for, say, >$25/mo?) Why, that would cut into the "bottom line", by gosh! To think PSI would expect to make a *profit* from this venture! Shocking, isn't it? >In PSI offered a complete "non-restrictive" news/mail feed, for slightly >more money (like, $100/mo) to cover the inbound "non-leaf-site" mail/news, >this would probably be acceptable to most customers. I knew they couldn't possibly afford to feed an entire city for $75/month, but I'm *sure* they could afford to do it for $100/month. -- John W. Temples -- john@jwt.UUCP (uunet!jwt!john)
jpr@jpradley.jpr.com (Jean-Pierre Radley) (03/14/91)
In article <2517@sun13.scri.fsu.edu> murray@sun13.scri.fsu.edu (John Murray) writes: >I had no opinion on this issue until one person mentioned what the cost for an >unrestricted site was. (BTW, can someone verify the $825-875 quarterly >cost for a full site?) Now I believe that this move is strictly a *marketing* >based decision on the part of PSI. I don't understand it on marketing grounds. Marty Schoffstall, earlier in this thread, pointed out how little additional burden was placed on PSI's facilities by handling mail and netnews on top of their commercial traffic. Seems to me that if you have something that costs little, and you can sell it thousands of times for barely any incremental cost for each sale, why wouldn't the marketing decision be to go for the highest possible volume? If they were to abandon their restatement of a contract to include terms, which as Sean just pointed out, were not at all in their original contracts, let alone in their brochures, and were to allow unimpeded and uncircumcsribed traffic, not only would they gain lots more new customers (maybe even some of those who have stated here that they would not now sign up with PSI), but who knows: some of those "third-party" sites that they now propose to shut out might grow up to be direct customers of PSI themselves. Jean-Pierre Radley NYC Public Unix jpr@jpradley.jpr.com CIS: 72160,1341
jerry@olivey.ATC.Olivetti.Com (Jerry Aguirre) (03/14/91)
In article <Dccqy4w163w@phoenix.com> stanley@phoenix.com (John Stanley) writes: >mra@searchtech.com (Michael Almond) writes: > Now PSI is saying (to some) that you may neither send nor receive >mail through PSI involving third parties. Now what use is a domain name? >If PSI did not intend to allow third party mail to be sent or recieved >through uupsi, why was getting a domain name one of the features of >their service? Why do they MX, if all they will allow is mail from PSI? Let me preface this by saying that I have not seen the contract nor do I have any relationship with PSI. The term "third party" seems to be used in a strange way here. When I think of "third party" as applied to the telephone or postal mail it means someone other than the originator or intended recipiant. By example if you make a phone call then you can record it. If you receive a phone call then you can record it. If you are making a phone call to someone else and I record it then I am breaking the law. You are number 1, :-) the person you are calling is number 2, :-) and I would be the 3rd party. The phone company is the carrier, not one of the parties. You are not restricted to placing calls only to the telephone company. Given that accepted usage, if you, as a PSI customer, send mail to someone external, or someone else external sends mail to you, then it is not third party mail. PSI does not count as one of the parties (unless the mail is to or from them). They are the carrier. There are a lot of regulations in fields from ham radio to the post office and they all assume this model of usage. On the other hand, PSI's refusal to accept incoming news postings sounds like a hasty decision, made without considering the ramifications. Companies do this all the time, alienating their customers for no benefit to themselves. They feel so threatened by the unexpected negative response that by the time they finally understand why it was a bad decision they have become stubborn and refuse to admit it. Usenet is not mail. There is no such thing as "third party" for usenet. Lumping usenet under the same restrictions as mail was ill considered. PSI sells news feeds, not posting privileges. You can not reasonably charge someone for posting if you are going to freely accept the same article when it comes in later via NNTP. As has been pointed out someone can easily circumvent the restrictions by making an inexpensive 5 minute call each night to upload the local postings to a backbone system. PSI will then attempt to feed, at no extra charge, the same article they wanted to charge money for receiving. The cost savings PSI offers is for the incoming feed, not the few postings the local cluster might generate.
schoff@uu.psi.com (Martin Schoffstall) (03/15/91)
>I don't understand it on marketing grounds. Marty Schoffstall, earlier in >this thread, pointed out how little additional burden was placed on PSI's >facilities by handling mail and netnews on top of their commercial traffic. >Seems to me that if you have something that costs little, and you can sell >it thousands of times for barely any incremental cost for each sale, why >wouldn't the marketing decision be to go for the highest possible volume? Let me try again. Backbone t1 bandwidth is available, especially in the evening, that is why we give our SCS+CCS customers NNTP news feeds for free. What is in short supply (has incremental cost) are $600 modems and $25/mo POTS lines to feed news through the normal phoneline/UUCP means. > >If they were to abandon their restatement of a contract to include terms, >which as Sean just pointed out, were not at all in their original contracts, >let alone in their brochures, and were to allow unimpeded and uncircumcsribed >traffic, not only would they gain lots more new customers (maybe even some of >those who have stated here that they would not now sign up with PSI), but who >knows: some of those "third-party" sites that they now propose to shut out >might grow up to be direct customers of PSI themselves. A couple of things here.. PSI goal is not to play to the current USENET "community" but to bring in a new group of participants who are interested in the service provision of a reliable accountable local leaf connection. (But if you want 3rd party you can have that at a higher price). In addition we've seen lots of "abuse" of the standards of practice in things like domain names and other issues with third parties, for instance people taking a .COM domain of theirs and handing it out to lots of other "organizations". This is a violation of the administrative law of the Internet. And let me take a minute to define "third party" which while well understood in the Internet world clearly isn't in the UUCP/Mail and USENET world. UUPSI provides email gateway service to our contracted customers anywhere they want to go on the Internet, Bitnet, etc. What a third party restiction represents is someone from "behind" our customer using our bandwith to get to those same places. To those organization we say get your own contract/connection to UUPSI. While I know that these positions are controversial, this was known in advance, it did not take me by surprise that there would be a response, unfortunately it was a particurally uniformed response due to the inaccuracy of the initial postings. PSI is simply using a different model of how to provide service, this includes the balancing of fees, contracts, technology, customer service, and access, it is not particurally evil, it is different. Marty
mjs@cbnews.att.com (martin.j.shannon) (03/15/91)
In article <1991Mar14.170247.10965@uu.psi.com> schoff@uu.psi.com (Martin Schoffstall) writes:
@ Let me try again. Backbone t1 bandwidth is available, especially
@ in the evening, that is why we give our SCS+CCS customers NNTP
@ news feeds for free. What is in short supply (has incremental
@ cost) are $600 modems and $25/mo POTS lines to feed news through
@ the normal phoneline/UUCP means.
This certainly sounds reasonable to me. But, see below.
@ A couple of things here.. PSI goal is not to play to the current
@ USENET "community" but to bring in a new group of participants who
@ are interested in the service provision of a reliable accountable
@ local leaf connection. (But if you want 3rd party you can have that
@ at a higher price).
Hmmm. Sounds like a little bit of "foot shooting" to me, but perhaps
"they" know what they're doing. I'm not sure I have any idea who might
fall into that "new group".... But, I'd love to hear about this new
group: give some examples of who these new folks might be.
@ UUPSI provides email gateway service to our contracted customers anywhere
@ they want to go on the Internet, Bitnet, etc. What a third party restiction
@ represents is someone from "behind" our customer using our bandwith to
@ get to those same places. To those organization we say get your own
@ contract/connection to UUPSI.
And here's the rub! It *seems* that if more folks do get thir own
connections to UUPSI, then UUPSI is artificially increasing the load on
the already expensive modems and POTS service. I have a feed that has
his modems at 90+% usage (off-peak, the only time I can afford to call
his machine). Is UUPSI going to certify a certain utilization maximum
(above which they will automatically add modems)?
@ While I know that these positions are controversial, this was known
@ in advance, it did not take me by surprise that there would be
@ a response, unfortunately it was a particurally uniformed response due to the
@ inaccuracy of the initial postings.
@
@ PSI is simply using a different model of how to provide service, this
@ includes the balancing of fees, contracts, technology, customer
@ service, and access, it is not particurally evil, it is different.
@
@ Marty
Marty, the only controversy *I* see is the denial of service to your
contracted customers. You've denied them the ability to forward *any*
mail they receive from anywhere outside UUPSI. Do they deny UUPSI the
ability to forward any mail *through* UUPSI's contracted customers?
(They should!) Does UUPSI *guarantee* that they will never attempt to
send mail/news *through* a customer's machine -- even if it is the
cheapest route?
Disclaimer: I do not currently use UUPSI. My interest is to be able to
afford a reasonable *personal* Internet connection -- some day (note: I
carry as full a news feed as I can get on my *personal* machine at
home -- out of my own pocket). I don't see UUPSI being a provider of
that sort of service with the current offerings that I've been made
aware of in this thread.
--
Marty Shannon; AT&T Bell Labs; Liberty Corner, NJ, USA
(Affiliation is given for identification only:
I don't speak for them; they don't speak for me.)
peter@ficc.ferranti.com (Peter da Silva) (03/15/91)
In article <2517@sun13.scri.fsu.edu> murray@sun13.scri.fsu.edu (John Murray) writes: > Did you miss the post that mentioned how much this other service cost? > First, think about this: In a newsfeed, what's more costly, providing the > downlink-side (the rest of the world) of a full or sizeable feed to a > leaf site or (relatively) small cluster of sites, or providing the > uplink side for all the posts that originate from that site or sites? To UUPSI, the cost for both is pretty much fixed. They're running a leased line into each area, hoping to find enough sites in that area to buy the service to pay for it. Up to the capacity of the leased line, there is no difference to them how much traffic each site generates. You're applying the logic of a variable-cost service to a fixed cost one. I presume that the pricing is designed so a full line feeding all the $75 sites it can handle will make a nice profit. If they only have a couple of such sites they lose money. They have to kick off freeloaders and force them to get their own lines or do without if they want to make a profit. I just wish they'd hurry up on the Houston POP. -- Peter da Silva. `-_-' peter@ferranti.com +1 713 274 5180. 'U` "Have you hugged your wolf today?"
peter@ficc.ferranti.com (Peter da Silva) (03/15/91)
In article <1991Mar14.052623.26604@jpradley.jpr.com> jpr@jpradley.jpr.com (Jean-Pierre Radley) writes: > Seems to me that if you have something that costs little, and you can sell > it thousands of times for barely any incremental cost for each sale, why > wouldn't the marketing decision be to go for the highest possible volume? That's what they're doing. Going for the highest possible volume. > [if UUPSI] were to allow unimpeded and uncircumcsribed > traffic, not only would they gain lots more new customers No, they wouldn't. Why should you buy from UUPSI if you can get the same product second-hand, not even shopworn? -- Peter da Silva. `-_-' peter@ferranti.com +1 713 274 5180. 'U` "Have you hugged your wolf today?"
peter@ficc.ferranti.com (Peter da Silva) (03/15/91)
In article <50462@olivea.atc.olivetti.com> jerry@olivey.ATC.Olivetti.Com (Jerry Aguirre) writes: > The term "third party" seems to be used in a strange way here. When I > think of "third party" as applied to the telephone or postal mail it > means someone other than the originator or intended recipiant. In this case it means a third party not mentioned on the original contract. I have a question... are people with contract paid up beyond the 45 days getting the letter? Now that might be a problem. -- Peter da Silva. `-_-' peter@ferranti.com +1 713 274 5180. 'U` "Have you hugged your wolf today?"
peter@taronga.hackercorp.com (Peter da Silva) (03/18/91)
stanley@phoenix.com (John Stanley) writes: > Now PSI is saying (to some) that you may neither send nor receive > mail through PSI involving third parties. Now what use is a domain name? That seems to be a misunderstanding. The intent is that mail should not be to or from some site on the other side of a subscriber. Perhaps PSI needs to make this clearer. -- (peter@taronga.uucp.ferranti.com) `-_-' 'U`
jpr@jpradley.jpr.com (Jean-Pierre Radley) (03/18/91)
In article <1991Mar14.170247.10965@uu.psi.com> schoff@uu.psi.com (Martin Schoffstall) writes: >In addition we've seen lots of "abuse" of the standards of practice >in things like domain names and other issues with third parties, >for instance people taking a .COM domain of theirs and handing it >out to lots of other "organizations". This is a violation of >the administrative law of the Internet. Could you expand on that a bit? Given some "place.com", what, or who, is or is not permitted by the "laws of the Internet" to be within that domain? Jean-Pierre Radley NYC Public Unix jpr@jpradley.jpr.com CIS: 72160,1341
schoff@uu.psi.com (Martin Schoffstall) (03/19/91)
> >Could you expand on that a bit? Given some "place.com", what, or who, >is or is not permitted by the "laws of the Internet" to be within that >domain? > As best I can....... It is assumed that "place" is a registered corporation, limited partnership, or some such for-profit entitity regulated inside some jurisdiction of the US (like a State, commonwealth, or territory). This organization is expected to generate names under its registered domain "place.com". No other organization or individual is to generate names with this domain. But there are exceptions, if you give accounts on your machine to other people, then you are fine. However it is verboten for other organizations to use your domain on their machines. To clear up some of the more interesting service holes there are other provisions; What is provided to deal with individual's machines is the .US domain. Some "organized" non-profit associations appear to be able to use the .org domain for distributed machine networks, (fidonet appears to do this), though they should speak for themselves. Rules have "changed", at one point in time organizations used to be placed under the .NET domain, that has now stopped. Marty
tneff@bfmny0.BFM.COM (Tom Neff) (03/19/91)
In article <7H31y1w163w@phoenix.com> stanley@phoenix.com (John Stanley) writes: > As taronga.hackercorp.com is 'another organization' (neither PSI nor >mine), I am prohibited from accepting mail from you, as I must receive >it through UUPSI. However, UUPSI is supposedly selling me the ability to >receive mail from other organizations through them. The latter is not in >writing, though, and the former is. Which is correct? You are supposed to rely on PSI's good faith in enforcing, not what they had you sign and date in ink, but what they generally sort of "meant." Better not decide that what YOU "meant" is different from what you signed, though! That would be breach of contract. :-)
peter@taronga.hackercorp.com (Peter da Silva) (03/19/91)
stanley@phoenix.com (John Stanley) writes: > I wish those who know what the intent of PSI is would tell the rest > of us how they found that out! I asked them. -- (peter@taronga.uucp.ferranti.com) `-_-' 'U`
phil@ls.com (Phil Eschallier) (03/19/91)
In article <1991Mar19.020431.28067@jpradley.jpr.com> jpr@jpradley.jpr.com (Jean-Pierre Radley) writes: > >I don't have here any of the documents which define domains. Simple >question, net-folk: > >I've got a domain: which sites may or not be IN my domain, and by what >criteria? > > Jean-Pierre Radley NYC Public Unix jpr@jpradley.jpr.com CIS: 72160,1341 i have been passively reading this thread -- admittedly i am somewhat confused. about 2 months ago i decided to get information from psi about their uucp services, hoping for a cheaper way to handle e-mail. in all, the information i got looked quite appealing and i decided i would go with their service (the $75/mo service that i gather is the service being discussed here). being short on time i didn't get to it right away -- but john eldrige did give me a follow up call, which i figured was convenient. i told him i was interested in the service and confirmed there were modems available in the philadelphia area. before we were about to cemment the deal, it (luckily) came out that we could not receive any mail other than what is addressed for .ls.com. that was the end of the conversation. who was he to tell me who/what mail my computers could pass. i understand theirs is a flat rate service -- if they did not have this restriction then i could buy a uucp account from psi then distribute e-mail to the entire philadelphia area -- they would have the expense of maintaining the equipement and would only take in $75/mo for the area. if your site is a leaf node, psi sounds like a good way to get connected. otherwise i would suggest a alternative service -- usually you get what you pay for. now i understand that psi is saying you can only send AND receive mail for your domain -- this may be appropriate for corporation with branches around the country. but for the leaf node (or one site domain), you are now paying $75/mo to be connected with yourself -- not exactly my definition of con- nectivity!! after much ramblings about my psi confusion -- the question at hand: what sites can be in a domain?? .ls.com was formed in the days of stargate and at that time it was a fee of $150/yr (i think ... it has been some time since i thought about stargate) for a second level domain -- but it was yours. you controlled routing, site memberships, domain parks ... the works. i still treat .ls.com in this fasion although there is no fee as i am no a uunet subsciber. but i'd be interested in knowing what the current rules are -- and if they have tightened (or been restricted), who's going to enforce them?? psi?? if so, get a different service. -- Phil Eschallier | E-Mail to: US Mail to: | INET: phil@ls.com 248B Union Street Lagniappe Systems | UUCP: ...!uunet!lgnp1!phil Doylestown, PA 18901 Computer Services | CIS: 71076,1576 VOICE: +1 215 348 9721
emv@ox.com (Ed Vielmetti) (03/19/91)
the solution would seem to be - rewrite headers to hide people in your "domain" - register as a .org, e.g. "*.emv.org" "friends of ed vielmetti" rather than a .com if you are indeed registering a domain for convenience and not for commercial traffic per se. (the name is actually quite arbitrary and .com sites are not subject to any different restrictions that .org sites face). - insist that your feed site (whatever it is) rewrite your headers to your specifications, within reason (e.g. not explicitly percentifying you behind their main relay) - do comparison shopping; find several possible places to get a connection from, and compare rates/services/etc. it would be nice to have a standard terms and conditions contract that multiple providers of this sort of service (news, mail, mx) would adhere to; that would make comparison shopping easier. a service providers scorecard would be interesting too. (but don't ask me to volunteer to fill it in, I'm an interested party) -- Msen Edward Vielmetti /|--- moderator, comp.archives emv@msen.com
stanley@phoenix.com (John Stanley) (03/19/91)
peter@taronga.hackercorp.com (Peter da Silva) writes: > stanley@phoenix.com (John Stanley) writes: > > I wish those who know what the intent of PSI is would tell the rest > > of us how they found that out! > > I asked them. And they sent me a letter which seems to contradict what they told you. Which bears more weight -- what they say to you or what they write in their contracts?
peter@taronga.hackercorp.com (Peter da Silva) (03/20/91)
jpr@jpradley.jpr.com (Jean-Pierre Radley) writes: > I've got a domain: which sites may or not be IN my domain, and by what > criteria? This is like the "what is a business" discussions people have about phone companies and BBSes. Some insist that if the IRS says it's a business, it is. The basic problem is that PSInet and the NIC have different requirements: PSInet is selling service to a "single site", and is regarding a single site as either (a) a single location, or (b) a single machine. They need to clarify this point (one hopes that they mean a single site, since there are over 40 machines on the premises at "ferranti.com"), and remove the confusing language about domains. Domains are an administrative convention not necessarily related to sites. -- (peter@taronga.uucp.ferranti.com) `-_-' 'U`
peter@taronga.hackercorp.com (Peter da Silva) (03/20/91)
stanley@phoenix.com (John Stanley) writes: > peter@taronga.hackercorp.com (Peter da Silva) writes: > > stanley@phoenix.com (John Stanley) writes: > > > I wish those who know what the intent of PSI is would tell the rest > > > of us how they found that out! > > I asked them. > And they sent me a letter which seems to contradict what they told you. Only if you choose to read the letter like a computer program. Hey, did you hear about the computer programmer who died in the shower? The instructions read "lather, rinse, repeat"... and he died of a stack overflow. > Which bears more weight -- what they say to you or what they write > in their contracts? That's up to you to decide. You always have the option of not doing business with them. I'm reserving judgement. -- (peter@taronga.uucp.ferranti.com) `-_-' 'U`
tneff@bfmny0.BFM.COM (Tom Neff) (03/20/91)
In article <TTG2RDB@taronga.hackercorp.com> peter@taronga.hackercorp.com (Peter da Silva) writes: >The basic problem is that PSInet and the NIC have different requirements: >PSInet is selling service to a "single site", and is regarding a single >site as either (a) a single location, or (b) a single machine. They need >to clarify this point (one hopes that they mean a single site, since >there are over 40 machines on the premises at "ferranti.com"), and remove >the confusing language about domains. Domains are an administrative >convention not necessarily related to sites. If PSI refuses to support true RFC1034 domains, they should not falsely register cut-rate leaf sites with the NIC as full domains. It is ludicrous to go ahead and post a wildcard record, then browbeat the customer for having the audacity to use it. The irony is that by aggressively segregating traffic to these cut-rate leaf sites, PSI not only ensures maximum wastage of its own disk and telecommunications resources per users served (thus shoving the population/performance dropoff point as far forward as possible), but also keeps the alternative option -- geographically proximate sites banding together to share the costs of a nonrestrictive joint feed from some other supplier -- perpetually attractive.
horne-scott@cs.yale.edu (Scott Horne) (03/21/91)
In article <5XG2Z8C@taronga.hackercorp.com> peter@taronga.hackercorp.com (Peter da Silva) writes:
<
<Hey, did
<you hear about the computer programmer who died in the shower? The
<instructions read "lather, rinse, repeat"... and he died of a stack
<overflow.
He needs a compiler which supports tail recursion. :-)
--Scott
--
Scott Horne ...!{harvard,cmcl2,decvax}!yale!horne
horne@cs.Yale.edu SnailMail: Box 7196 Yale Station, New Haven, CT 06520
203 436-1817 Residence: Rm 1817 Silliman College, Yale Univ
"Pi4 nai3 ren2 shen1 zhi1 qi4, qi3 you3 bu2 fang4 zhi1 li3." --Mao Zedong
xanthian@zorch.SF-Bay.ORG (Kent Paul Dolan) (03/21/91)
jpr@jpradley.jpr.com (Jean-Pierre Radley) writes: > When I first subscribed to the $75/mo service, one of the first > telephone conversations they had with me concerned what I wished for a > domain name, which they obtained for me quite efficiently and rapidly. > Part of their objections to my current use of their service involves > what PSI says is inappropriate adherence to my domain. > If you don't recall from earlier in this thread, John Eldredge is > PSI's Director of Sales & Marketing. In a letter just now received, he > writes: > "[you're paid through 5/8/91, etc.] "It is also clear that you are > including other people's machines under your domain which is not > consistent with the use of the '.com' domain." > After suggesting that I might want to consider switching to the UUPSI > Redistribution Service, at a higher tariff, he continues: > "[re-execute new verion of old contract, or sign the higher-rate > contract, for service beyond 4/15/91, etc.] "However, if the current > domain issue is not cleared up, we may have to settle this situation > sooner." > I don't have here any of the documents which define domains. Simple > question, net-folk: > I've got a domain: which sites may or not be IN my domain, and by what > criteria? Sure does sound like, under the accepted definition of "domain", UUPSI has indeed, as mentioned in an earlier posting, pulled a classic "bait and switch" confidence game on you. This is highly illegal, and district prosecuting attorneys love these cases; because they're a lead pipe cinch to prosecute (judges _detest_ business scams); look like one on the side of the angels, defending the little guy; and up the "win" column score for the guy/gal handling the case. Keep every scrap of paper, record your phone conversations, and save your email and any replies to your news postings from UUPSI representatives. If you have the time for the court appearances, you can have a lot of fun with this one. Kent, the man from xanth. <xanthian@Zorch.SF-Bay.ORG> <xanthian@well.sf.ca.us>
fitz@wang.com (Tom Fitzgerald) (03/22/91)
>> Could you expand on that a bit? Given some "place.com", what, or who, >> is or is not permitted by the "laws of the Internet" to be within that >> domain? schoff@uu.psi.com (Martin Schoffstall) writes: > It is assumed that "place" is a registered corporation, limited partnership, > or some such for-profit entitity regulated inside some jurisdiction of the US > (like a State, commonwealth, or territory). This organization is expected > to generate names under its registered domain "place.com". No other > organization or individual is to generate names with this domain. Who or what assumes this? Is this assumption actually written down anywhere? There are currently whole domain parks and regional organizations that have .com domains. This wasn't a problem when they were first created, and I haven't seen anything announced since then that implies it is has become illegal. Currently, it looks to me like the *.mn.org, *.mv.com and *.assabet.com domains exist specifically to share a domain, and that's it. --- Tom Fitzgerald Wang Labs fitz@wang.com 1-508-967-5278 Lowell MA, USA ...!uunet!wang!fitz
emv@ox.com (Ed Vielmetti) (03/22/91)
In article <b2ng7e.hkc@wang.com> fitz@wang.com (Tom Fitzgerald) writes: schoff@uu.psi.com (Martin Schoffstall) writes: > It is assumed that "place" is a registered corporation, limited partnership, > or some such for-profit entitity regulated inside some jurisdiction of the US > (like a State, commonwealth, or territory). This organization is expected > to generate names under its registered domain "place.com". No other > organization or individual is to generate names with this domain. not true. my internet managers phonebook records a "pcs.com" which is in Germany. I believe that the NIC will not refuse to register anyone in .com as long as they fill out the forms right. They will refuse to register for-profit firms in ".org" and non-degree-granting institutions in ".edu"; I think you need to jump through special hoops to be a .net. heck, there's even Canadian and German .edu sites (free trade agreements at work). Currently, it looks to me like the *.mn.org, *.mv.com and *.assabet.com domains exist specifically to share a domain, and that's it. I also see .wa.com operating in this manner; looks like a * MX record pointing to UUNET. Which means no doubt that they are paying by the hour rather than by the month for their service ... -- Msen Edward Vielmetti /|--- moderator, comp.archives emv@msen.com
dennis@gpu.utcs.utoronto.ca (Dennis Ferguson) (03/22/91)
In article <EMV.91Mar21155151@poe.aa.ox.com> emv@ox.com (Ed Vielmetti) writes: >schoff@uu.psi.com (Martin Schoffstall) writes: >> It is assumed that "place" is a registered corporation, limited partnership, >> or some such for-profit entitity regulated inside some jurisdiction of the US >> (like a State, commonwealth, or territory). This organization is expected >> to generate names under its registered domain "place.com". No other >> organization or individual is to generate names with this domain. > > not true. my internet managers phonebook records a "pcs.com" which is > in Germany. I believe that the NIC will not refuse to register anyone > in .com as long as they fill out the forms right. They will refuse I think the last sentence is true. For further supporting examples, try newbridge.com, sq.com, clarinet.com, sobeco.com and geac.com (this list is hardly complete, just the ones that come to mind). I also don't think I've ever seen constraints concerning what may be named using the domain name documented anywhere, indeed the quote above is the first I've heard of it. References would be nice. Dennis Ferguson
rissa@mejac.palo-alto.ca.us (Patricia O Tuama) (03/22/91)
In article <1991Mar21.054111.5962@zorch.SF-Bay.ORG> xanthian@zorch.SF-Bay.ORG (Kent Paul Dolan) writes: >This is highly illegal, and district prosecuting attorneys love these >cases; because they're a lead pipe cinch to prosecute (judges _detest_ >business scams); look like one on the side of the angels, defending the >little guy; and up the "win" column score for the guy/gal handling the >case. Over time, he enters more and more kill files, and really has created his own private little newsgroup-for-one, in, but not a part of, talk.bizarre, in which his incredible cleverness, wonderful sense of humor, and great creative writing skill go completely unnoticed except by him; really, not that much of a change when you think about it. Kent, the man from xanth. <xanthian@Zorch.SF-Bay.ORG> <xanthian@well.sf.ca.us> -- Of course, in a sense, I have partially accomplished that for myself throughout a large part of the net. Oh, well, another day, another interminable posting from Casa Xanth. K*nt Dolan <1991Mar21.070429.7540@zorch.SF-Bay.ORG>
peter@ficc.ferranti.com (Peter da Silva) (03/22/91)
In article <73552093@bfmny0.BFM.COM> tneff@bfmny0.BFM.COM (Tom Neff) writes: > If PSI refuses to support true RFC1034 domains, they should not falsely > register cut-rate leaf sites with the NIC as full domains. It is > ludicrous to go ahead and post a wildcard record, then browbeat the > customer for having the audacity to use it. You're right, they shouldn't use wildcards. If they are this is highly bogus. It's easy enough to set up separate MXes for every site in a domain. > [UUPSI] keeps the alternative option -- geographically proximate sites > banding together to share the costs of a nonrestrictive joint feed from some > other supplier -- perpetually attractive. That's true, and I believe they're fully aware of that. They depend on the low front-end cost making this option less desirable than a fixed-rate leaf feed. For that matter, that "other supplier" might still be UUPSI... their higher priced services are still in the ballpark with UUNET. -- Peter da Silva. `-_-' peter@ferranti.com +1 713 274 5180. 'U` "Have you hugged your wolf today?"
schoff@uu.psi.com (Martin Schoffstall) (03/23/91)
> not true. my internet managers phonebook records a "pcs.com" which is > in Germany. I believe that the NIC will not refuse to register anyone > in .com as long as they fill out the forms right. They will refuse > to register for-profit firms in ".org" and non-degree-granting > institutions in ".edu"; I think you need to jump through special hoops > to be a .net. heck, there's even Canadian and German .edu sites (free > trade agreements at work). My impression/belief-system is that (1) there are some grandfather situations (2) there have been some mistakes (3) that the issue of multinational corporations makes a number of foreign sites be placed under .COM [(3) is a big problem by the way, at this week's NorthAmerican Directory Forum, a rather large company finally agreed that the naming scheme for X.500 should not include a special place for multinationals, but a form of "symbolic links" under the ISO nation prefixes (like @c=US) would be "the standard"] > > Currently, it looks to me like the *.mn.org, *.mv.com and *.assabet.com > domains exist specifically to share a domain, and that's it. Yep. There are .ORG ones and that is probably where they belong, the .COM ones are "mistakes". Just as background to this, things are becoming even more fluid as the long time contract support that SRI did for domains and all the other good things has now been won by a DC based organization. Marty
emv@ox.com (Ed Vielmetti) (03/23/91)
In article <1991Mar22.174928.26395@uu.psi.com> schoff@uu.psi.com (Martin Schoffstall) writes:
Just as background to this, things are becoming even more fluid
as the long time contract support that SRI did for domains and
all the other good things has now been won by a DC based organization.
who won the contract?
--
Msen Edward Vielmetti
/|--- moderator, comp.archives
emv@msen.com
schoff@uu.psi.com (Martin Schoffstall) (03/23/91)
> Just as background to this, things are becoming even more fluid > as the long time contract support that SRI did for domains and > all the other good things has now been won by a DC based organization. > >who won the contract? > Someone should pull this from CBD, I have this piece of info 3rd hand: - Network Information Systems - Reston, VA - "partially owned" by MCI and Infonet I'd put this at 33% accuracy. I'll find their address so you can send them your resume, they may not know how to anon ftp and post poscript files! ;-) Marty
jpp@tygra.UUCP (John Palmer) (03/23/91)
Regarding domain name service: Are there any sites which provide this to .COM sites (other than UUNET and UUPSI)?? Even a slight fee would be no problem, just as long as these restrictions didn't apply. One thing about PSI that is a big drawback is the fact that you cannot get an IP address with their HOST-DCS service. Each time you dial into one of their POP's, you get assigned a temporary IP address. The result: No one can TELNET/FTP to your site. It seems to be a trivial problem for them to reprogram their routers to know what POP each of their customers is assigned to and send any queries/connection requests to that POP when a connection is requested to one of their customers. The worst that can happen is: NOT RESPONDING - TIMED OUT. John Palmer -- CAT-TALK Conferencing System | "Buster Bunny is an abused | E-MAIL: +1 313 343 0800 (USR HST) | child. Trust me - I'm a | cat@tygra.UUCP +1 313 343 2925 (TELEBIT PEP) | professional..." | ..sharkey!tygra! ********EIGHT NODES*********** | -- Roger Rabbit | ..cat
pda@Dixie.Com (Paul D. Anderson) (03/24/91)
phil@ls.com (Phil Eschallier) writes: >if your site is a leaf node, psi sounds like a good way to get connected. >otherwise i would suggest a alternative service -- usually you get what >you pay for. This is not directed at Phil - but to the group. Ultimately, you have to ask, "What good is PSInet or ALTERnet or any of these nets, if they don't let us handle 3rd party mail?" Now - I know that PSI has a dual rate scheme such that you can handle 3rd party stuff, too, but let's gloss over that for a moment, since a lot of folks can't pay that rate scale yet... The only VALUE in connecting to a network is to give you access to more individuals (if you're into exchanging info with peers around the world). If PSInet restricts how much data flows across the net, *where* *is* *the* *value* *added*. All I will have bought is an expensive long distance telephone connection that I can only talk to a few neighbors on. If a network has a policy which will restrict how many people I can talk to, then I'm not sure it's of benefit (Just like Sprint was worthless to me when they only had POP's in major cities). >now i understand that psi is saying you can only send AND receive mail for >your domain -- this may be appropriate for corporation with branches around >the country. but for the leaf node (or one site domain), you are now paying >$75/mo to be connected with yourself -- not exactly my definition of con- >nectivity!! Exactly. I guess demand mode dialing to any city via uucp and having a system there route to the appropriate site is much more beneficial than paying a flat rate to not be able to talk to anyone that I need to talk to. It certainly will remain more cost effective than connecting to a commercial net, at this point. Hey, Phil - wanna be one of the NEWnet(tm) end points? I'll be one here in Atlanta. 1/2 :-) -paul -- * Paul Anderson * Dixie Communications * (404) 578-9547 * pda@dixie.com *
schoff@uu.psi.com (Martin Schoffstall) (03/26/91)
>One thing about PSI that is a big drawback is the fact that you cannot get >an IP address with their HOST-DCS service. Each time you dial into one of >their POP's, you get assigned a temporary IP address. This is ONE of the services that we offer, other services such as CCS and SCS do not have this "limitation". >The result: No one can >TELNET/FTP to your site. It seems to be a trivial problem for them to >reprogram their routers to know what POP each of their customers is assigned >to and send any queries/connection requests to that POP when a connection is >requested to one of their customers. The worst that can happen is: NOT >RESPONDING - TIMED OUT. I'm not sure if cisco/proteon/wellfleet/nsc would agree that it is trivial to reprogram their routers; however, on-demand Internet access is not trivial either. There are questions such as (1) is "on-demand" address assignment a good thing? From looking at the PPP spec and some implementations it appears that the answer is yes. It is probably a good thing from the perspective of the Internet Address space also, right now the only granularity of long term official address assignment is 8 bits (a whole class C network number). (2) is "on-demand" client-only access to the Internet useful, (HOST-DCS)? I think the answer is yes. There appear to be 100,000's of Internet hosts which are client-only today. With Intercon/FTPSoftware/etc support POP3 et al of the standard uses (ftp/telnet/mail) are provided. (3) is there going to be changes and evolution in the future? Absolutely, the Internet, internetworking products, etc are rapidly evolving. Marty
AMillar@cup.portal.com (Alan DI Millar) (03/27/91)
>Ultimately, you have to ask, "What good is PSInet or ALTERnet or any of >these nets, if they don't let us handle 3rd party mail?" > >>now i understand that psi is saying you can only send AND receive mail for >>your domain Wait a minute. This is getting ludicrous. I'm not a customer of PSI, and I may not like some of the things I am hearing about them, but they aren't stupid. You aren't paying just to talk to yourself. "Send AND receive mail for your domain" means "send mail from" and "receive mail for" your domain. Picture the following: SiteA -----------+----------- SiteB | PSI / \ / \ You SiteC / SiteD You are a customer of PSI. So is SiteC. SiteA and SiteB are "the outside world". SiteD is a another machine connected to just you. What is implied above is that the only one you can talk to with the PSI "leaf contract" is SiteC. Give me a break. Of course you can talk to SiteA and SiteB and _the_rest_of_the_world_. What PSI is saying is that they won't accept traffic through you between --==>> SiteD <<==-- and the rest of the world. If you want to send mail from SiteD to SiteA, it has to be through another route than PSI. Now, you still may not like *that* restriction, but let's not grossly exagerate it. With that said, I think I've had about enough of this thread. :-) - Alan Millar flames to: AMillar@cup.portal.com
AMillar@cup.portal.com (Alan DI Millar) (03/27/91)
>>The result: No one can >>TELNET/FTP to your site. It seems to be a trivial problem for them to >>reprogram their routers to know what POP each of their customers is assigned >>to and send any queries/connection requests to that POP when a connection is >>requested to one of their customers. The worst that can happen is: NOT >>RESPONDING - TIMED OUT. > >I'm not sure if cisco/proteon/wellfleet/nsc would agree that it is >trivial to reprogram their routers; however, on-demand Internet >access is not trivial either. Funny, I though trivial on-demand Internet access was the whole point of the dial-up IP code I recently saw on Usenet. The Telebit NetBlazer seems to think it's pretty trivial... (For those who haven't seen it, the NetBlazer is a dial-up IP router. It uses dial-up modems to establish IP connections using SLIP or PPP. I'm not affiliated with Telebit; I just think it's a neat product) - Alan Millar
schoff@uu.psi.com (Martin Schoffstall) (03/27/91)
>Funny, I though trivial on-demand Internet access was the whole point >of the dial-up IP code I recently saw on Usenet. The Telebit >NetBlazer seems to think it's pretty trivial... (For those who haven't >seen it, the NetBlazer is a dial-up IP router. It uses dial-up modems >to establish IP connections using SLIP or PPP. I'm not affiliated >with Telebit; I just think it's a neat product) > And PSINet has supported a form of dialup Internet access since the Fall of 1990 for a flat fee.... But let's think about providing this as a service: (1) who makes the phone calls? If it is the customer, then everything is fine for the service provider vis a vis billing the flat fee, but what about the site inside the leased line Internet who wants to reach that customer, he's out of luck. If they both make the phone calls then the service provider has to keep track of how many message units are consumed (hopefully through SNMP stats) multiply by the appropriate amount and bill it in addition to the fixed amount. (2) How would you like to configure the DNS? Is the primary at the customer site or inside the leased line Internet. (3) Where does email go for the customer site when it isn't connected? If the service provider MX's for that site, how does it make final delivery, UUCP? (too UNIX centric), POP3 (too individual centric), SMTP with TURN (not readily available), SMTP without TURN (you have to hope to match the window of availability), MMDF/TCP? In reality you want to do whatever the customer wants, so you have a lot of variability between customers. I believe that dialup/switched/on-demand Internet access is a great thing, we have over 100 customers who do this right now; however, it is not trivial (for the reasons that I've outlined, and many reasons that I haven't), and more importantly it is pointing out some additional work that needs to be done in the Internet suite to make it extrememly useful to the users. Marty
tneff@bfmny0.BFM.COM (Tom Neff) (03/27/91)
In article <40584@cup.portal.com> AMillar@cup.portal.com (Alan DI Millar) writes: >What is implied above is that the only one you can talk to with the PSI >"leaf contract" is SiteC. Give me a break. Of course you can talk to >SiteA and SiteB and _the_rest_of_the_world_. What PSI is saying is that >they won't accept traffic through you between --==>> SiteD <<==-- and >the rest of the world. That's what they're telling us now, informally on the net. It's not what you actually sign. What you sign says: "No electronic mail from another organization should be sent or received through UUPSI." Really! I only make a point of all this because... this is what it's going to be like, more and more, as capitalism hits the Net. You're going to see this odd mix of practices and personalities impacting things. The press release replaces the RFC. Profit margins, and strategies for maintaining them, begin to take precedence over basic connectivity and flexibility of the net. The founding net paradigm -- everyone investing a little of their resources in order to reap rewards of creativity and edification -- turns slowly inside out: information will now be doled out in restricted, metered driblets, for the direct enrichment of the supplier. That would never have built the Net we have today. It will take years to demolish it, but I advise readers to watch for the signs.
asp@uunet.UU.NET (Andrew Partan) (03/28/91)
In article <8640@rsiatl.Dixie.Com>, pda@Dixie.Com (Paul D. Anderson) writes: > Ultimately, you have to ask, "What good is PSInet or ALTERnet or any of > these nets, if they don't let us handle 3rd party mail?" Ummm... Hold on a minute here. We don't care what you do with your connection. If you have a UUCP connection to us & want to hook up a stack of other sites behind you & pass traffic every which way, fine. We don't care. If you have an AlterNet connection to us & want to hook up a pile of uucp sites behind you & pass traffic every which way, fine. We don't care. If you have an AlterNet connection to us & want to hook up a pile of sites via TCP/IP, talk to us. We are willing to deal. --asp@uunet.uu.net (Andrew Partan)
tneff@bfmny0.BFM.COM (Tom Neff) (03/28/91)
In article <-EBAZ+5@xds13.ferranti.com> peter@ficc.ferranti.com (Peter da Silva) writes: >In article <61034439@bfmny0.BFM.COM> tneff@bfmny0.BFM.COM (Tom Neff) writes: >> I only make a point of all this because... this is what it's going to be >> like, more and more, as capitalism hits the Net. You're going to see >> this odd mix of practices and personalities impacting things. > >But we've always had that: Domain versus bang paths, rabid rerouters >versus literalists, and so on. Look at the BITNET gateway stuff. Politics >is an unfortunately major pasttime on the net, as it is in any collection >of people. This is just more of the same. I'm afraid I can't agree. Yes, politics and controversy of one sort have been a recurrent fact of net administrative life for some time. But these have largely been of a technical or practical nature. What's new is the ETHICAL, motivational, even cultural aspect. Although the *technical* debate over rabid rerouting or heterogeneous network gatewaying can get pretty hot, at least we are not treated to transplanted sales types spewing Rotarian commonplaces about the way "the market is going." The common, underlying assumption of the Net I know is that moving as much data as possible to as many people as possible is the MAIN priority. The debate simply rages over the best way to do it. It used to be that no big net player would *demean* itself by squatting on a perfectly functional, standard, available resource -- technically completely ready to go -- and refuse to pass data because it was insufficiently lucrative. But that's what we're starting to see. And it's going to get worse before -- if -- it gets better. Usenet will have to choose between Babbage and Babbitt.
rfarris@rfengr.com (Rick Farris) (03/28/91)
In article <40584@cup.portal.com> AMillar@cup.portal.com (Alan DI Millar) writes: | SiteA -----------+----------- SiteB | | | PSI | / \ | / \ | You SiteC | / | SiteD | | | What PSI is saying is that they won't accept traffic through | you between --==>> SiteD <<==-- and the rest of the world. | If you want to send mail from SiteD to SiteA, it has to be | through another route than PSI. Now wait a minute, that doesn't make any sense. That would imply that PSI is only of use to connect to other PSI sites. *Surely* PSI means that _one_end_ of a mail route must terminate at a PSI site, musn't they? -- Rick Farris RF Engineering POB M Del Mar, CA 92014 voice (619) 259-6793 rfarris@rfengr.com ...!ucsd!serene!rfarris serenity bbs 259-7757
peter@ficc.ferranti.com (Peter da Silva) (03/28/91)
In article <61034439@bfmny0.BFM.COM> tneff@bfmny0.BFM.COM (Tom Neff) writes: > I only make a point of all this because... this is what it's going to be > like, more and more, as capitalism hits the Net. You're going to see > this odd mix of practices and personalities impacting things. But we've always had that: Domain versus bang paths, rabid rerouters versus literalists, and so on. Look at the BITNET gateway stuff. Politics is an unfortunately major pasttime on the net, as it is in any collection of people. This is just more of the same. -- Peter da Silva. `-_-' peter@ferranti.com +1 713 274 5180. 'U` "Have you hugged your wolf today?"
time@ice.com (Tim Endres) (03/29/91)
In article <61034439@bfmny0.BFM.COM>, tneff@bfmny0.BFM.COM (Tom Neff) writes: > I only make a point of all this because... this is what it's going to be > like, more and more, as capitalism hits the Net. You're going to see > this odd mix of practices and personalities impacting things. The press > release replaces the RFC. Profit margins, and strategies for > maintaining them, begin to take precedence over basic connectivity and > flexibility of the net. The founding net paradigm -- everyone investing > a little of their resources in order to reap rewards of creativity and > edification -- turns slowly inside out: information will now be doled > out in restricted, metered driblets, for the direct enrichment of the > supplier. That would never have built the Net we have today. It will > take years to demolish it, but I advise readers to watch for the signs. This may be too pessemistic of a view... Commercialism of the "Internet" is here. This is a very good thing for companies and independent professionals. A very good thing. Commercialism of "Usenet" is happening, but I don't believe that it will "squeeze-out" the traditional Usenet player. Commercial entities will never successfully force *everyone* to pay for their information, and I don't think they want to. Persons will still almost always find access to Usenet via UUCP from "cooperative" neighbors. Hell, I know of several companies just locally that will give a feed to anyone who asks (even though they are paying up to $1000/mo for their feed). If things go well, everyone will get what they want. Commercial entities "pay" for extremely reliable service and up-to-the-minute news and mail. Netters will still "network" for reasonable and timely news and mail. To insure this happens, we only need what we have had all along: -- Usenet users with the original vision. -- tim. ------------------------------------------------------------- Tim Endres | time@ice.com ICE Engineering | uupsi!ice.com!time 8840 Main Street | Voice FAX Whitmore Lake MI. 48189 | (313) 449 8288 (313) 449 9208
peter@ficc.ferranti.com (Peter da Silva) (03/30/91)
In article <12312497@bfmny0.BFM.COM> tneff@bfmny0.BFM.COM (Tom Neff) writes: > Yes, politics and controversy of one sort have been a recurrent fact of > net administrative life for some time. But these have largely been of a > technical or practical nature. What's new is the ETHICAL, motivational, > even cultural aspect. As soon as the Internet and Usenet joined the ethical, motivational, and cultural differences between the two groups of users and sites has been a problem. The whole debate over DNS versus the maps is a result of the differences between Usenet and Internet culture. An RFC blessing bang paths would have been the logical solution to the problem. The technical botch that resulted, with % and source-routing and the like was not the result of a technical problem. It was cultural. The creation of alt and biz as a place for stuff that doesn't belong on the Internet is a technical solution to an ethical problem. UUNET and UUPSI are themselves a result of this... not a cause. > at least we are not treated to > transplanted sales types spewing Rotarian commonplaces about the way > "the market is going." No, we're treated to the sight of the DNS folks preserving the empire of the atsign. Or look at TCP/IP versus OSI: you think that's not a matter of marketing on another level? For that matter the whole UNIX phenomenon is the result of marketing: BSD surely didn't win on the Internet as a result of technical superiority. It won because it was "free". > The common, underlying assumption of the Net I > know is that moving as much data as possible to as many people as > possible is the MAIN priority. You mean like alt.sex.pictures? > It used to be that no big net player would *demean* itself by squatting > on a perfectly functional, standard, available resource -- technically > completely ready to go -- and refuse to pass data because it was > insufficiently lucrative. But that's what we're starting to see. And > it's going to get worse before -- if -- it gets better. Just try to get a UUCP news feed from a typical university or business. You're not a small personal machine trying to stay on the net as your feeds come and go at the whim of random managers and administrators. This is just business as usual. At least UUPSI doesn't care about the color of your money and isn't subject to the whims of the state legislature. > Usenet will have to choose between Babbage and Babbitt. Babbage never did get his machine built. -- Peter da Silva. `-_-' peter@ferranti.com +1 713 274 5180. 'U` "Have you hugged your wolf today?"
stanley@phoenix.com (John Stanley) (04/01/91)
peter@ficc.ferranti.com (Peter da Silva) writes: > At least UUPSI doesn't care about the color of > your money and isn't subject to the whims of the state legislature. Oh, yes, UUPSI cares about the color of my money. If I tried to pay them with blue dollar bills they would certainly object.