- Test - lists.nanog.org

Re: links on the blink
by postel＠ISI.EDU 12 Nov '95

12 Nov '95

> The net's usability slowly declined over the years until reaching a > level even network engineers couldn't hand over, then it would > suddenly take a leap to a new bandwidth plateau. > How about a few positive suggestions which could be implemented in > the short-term. Sean: Get more bandwidth. --jon. Dennis: Right. We are way overdue for a major step up in bandwidth. What we need is OC12 bit pipes (no stinking ATM cells) and Routers that can handle them. --jon.

3 2

Re: Bell vs. Internet Providers (fwd)
by Roger Fajman 11 Nov '95

11 Nov '95

> Short: off-peak usage of telco facilities is not costing the telco any > real money, (this is especially true for LD companies!) as the capacity > planning (and purchase) is governed by peak usage, any screams by them in > that general direction should be ignored, their monopolistic actions > fought with a vengeance: EdTel of Edmonton,Alberta started making noises > as described no sooner than they were beginning to plan their own internet > service. I hear they have a special tariff discount for fax lines (business > or residential) too, or at least used to. > Are they talking with two tounges ? > > bye,Kai Bell Atlantic has recently been running an ad campaign here in the Washington, DC, area offering a $30 discount on the installation of second residential lines. The ads suggest that they are great for modem users and families with teenagers. I don't think they would do this if they didn't think they would make money on it. However, at the same time they are trying to introduce a residential ISDN tariff in Maryland that has all calls, both voice and data, charged by the minute, even during off peak times.

1 0

Re: Bell vs. Internet Providers (fwd)
by kai＠netcom.com 11 Nov '95

11 Nov '95

>ISP's should consider using a DS1 or better to deliver dialtone for their >modem banks. It makes alot of sense for the ISP and the phone company. >Delivering dialtone over a T1 is certainly more cost effective than paying >a %400 increase for Centrex lines. > Scott Nope, it isn't . At least not in Nynex land. Tariffs are stupid and don't make sense , I know. If you get your phone lines delivered via T1 in NYC, you pay about $550 for a local T1. 24 POTS lines run about $600 (business rate). Now of course, you are sitting there with your SmartJack, and need to shell out $3K-$10K for a channel bank, or buy a USR Total Control rack (or a Livingston Pop-in-a-rack). Your cost per line/modem/terminal server port has just exploded, congratulations, but you don't save any significant amount per month for this arrangement. Given that you have to provide about one line per 10 customers (SLIP/PPP) these days for decent quality service, you are in serious trouble with your cash flow...

2 1

Re: Bell vs. Internet Providers (fwd)
by Vadim Antonov 10 Nov '95

10 Nov '95

Sounds awfully familiar... The knee jerk reaction of Minsvyazi in USSR was exactly the same. Good thing, it later turned out that Anti-Monopoly Commitee were in fact customers of RELCOM. Thanks, Paul. --vadim Date: Fri, 10 Nov 1995 06:59:46 -0500 From: Paul Ferguson <pferguso(a)cisco.com> Subject: Bell vs. Internet Providers (fwd) >From the 'For What Its Worth' Department. - paul BELL VS. INTERNET PROVIDERS Bell Canada plants to increase line rate charges for Internet service providers by as much as 400%, saying the higher charge is necessary because those providers use lines for 55 to 60 minutes every hour compared with voice usage on a Centrex line of about 10 minutes. The service providers say the move aims to eliminate competition for Bell when it introduces its Worldlinx Internet access by the end of this year. (Montreal Gazette 9 Nov 95 D7) ***************************************************************** Edupage, 9 Nov 95. Edupage, a summary of news items on information technology, is provided three times each week as a service by Educom, a Washington, D.C.-based consortium of leading colleges and universities seeking to transform education through the use of information technology. *****************************************************************

1 0

Re: links on the blink (fwd)
by rcraig＠cisco.com 10 Nov '95

10 Nov '95

At 18:28 95/11/10, Vadim Antonov wrote: >Michael Dillon wrote: > >>By "load sharing" I presume you mean some sort of TDM where you have n >>real lines and every n'th packet goes out on any particular line. I >>suppose this would be even simpler to do at the byte level if we assume >>that all n lines go to the same endpoint. > >Yep. However the round-robin load-sharing you described breaks the >sequencing of packets (i.e. they will often arrive in the different >order from which they were sent) and that breaks a lot of TCP stacks >out there. Cisco's solution is to to do sharing on per-host or per-network >basis (i.e. all packets to host/network go along the same path) which >is, to say, of very little use in the backbone. It breaks down particularly >spectacularly when you have high degree of aggregation or "calling centers", >i.e. hosts or networks attracting particularly heavy traffic. > >I have a simple solution for the sequencing problem in load sharing >setups, which produces load patterns nearly as good as round-robin. >Don't ask how if you're not going to invest :) That algorithm is >central to the proposed hypercube router architecture, and i'm >past the age when i did things just for fun. > >--vadim Vadim! You've already said too much!! :_) Robert.

1 0

Re: links on the blink (fwd)
by Vadim Antonov 10 Nov '95

10 Nov '95

Michael Dillon wrote: >By "load sharing" I presume you mean some sort of TDM where you have n >real lines and every n'th packet goes out on any particular line. I >suppose this would be even simpler to do at the byte level if we assume >that all n lines go to the same endpoint. Yep. However the round-robin load-sharing you described breaks the sequencing of packets (i.e. they will often arrive in the different order from which they were sent) and that breaks a lot of TCP stacks out there. Cisco's solution is to to do sharing on per-host or per-network basis (i.e. all packets to host/network go along the same path) which is, to say, of very little use in the backbone. It breaks down particularly spectacularly when you have high degree of aggregation or "calling centers", i.e. hosts or networks attracting particularly heavy traffic. I have a simple solution for the sequencing problem in load sharing setups, which produces load patterns nearly as good as round-robin. Don't ask how if you're not going to invest :) That algorithm is central to the proposed hypercube router architecture, and i'm past the age when i did things just for fun. --vadim

1 0

Re: Bell vs. Internet Providers (fwd)
by Jeff.Ogden＠um.cc.umich.edu 10 Nov '95

10 Nov '95

The question is where are the real costs of a local phone call? Are they in the call setup? Are they in the length of time for the call? Some combination? What switch resources need to be expanded to deal with long holding time calls? What switch resources can be reduced if you need to process fewer call setups? It is possible that modem calls are a bad deal for phone companies, but it is also possible they are a good deal. Does anyone really know? -Jeff Ogden Merit

3 2

Re: Bell vs. Internet Providers (fwd)
by kai＠netcom.com 10 Nov '95

10 Nov '95

The question is where are the real costs of a local phone call? Are they in the call setup? Are they in the length of time for the call? Some combination? What switch resources need to be expanded to deal with long holding time calls? What switch resources can be reduced if you need to process fewer call setups? It is possible that modem calls are a bad deal for phone companies, but it is also possible they are a good deal. Does anyone really know? -Jeff Ogden Merit This is classic traffic engineering. The cost of providing service is almost exclusively governed by capacity size, and that is chosen following a sole parameter: peak usage. You never ever want to run out of local switch trunks for local calls: modern switch matrixes can interconnect 50% of all lines internally, interconnecting trunks to other switches can handle maybe 20% of line capacity. Traffic patterns show that peak usage is during business hours, with call volume (lines engaged) being 3-4 times the average of after-hour use. There is no telco in the world, that can show me they are being ripped off by the growing number of modem users, this is really just another case of 'cashing in on the super information highway' . The few people who stay on 24 hours are an EXCEPTION, not a rule, and in most cases such people have several phone lines in their home (I have 3, one is on 24 hours , hence on average I don't use my phone lines more than 33% of all times, as the other lines are seldomly in use: the phone co. is cashing in by selling me three lines !). Also: there are many people who get second phone lines exclusively for modem use now, hence a single modem line an ISP has installed might create an additional 2-3 lines with residential customers. Did you ever hear the phone co. complaining about too many residential (or other) phone lines ? All Nynex whines of how residential service would really cost them much more than the customer is actually paying aside: this must be one of the really dirty secrets of the RBOCS. Short: off-peak usage of telco facilities is not costing the telco any real money, (this is especially true for LD companies!) as the capacity planning (and purchase) is governed by peak usage, any screams by them in that general direction should be ignored, their monopolistic actions fought with a vengeance: EdTel of Edmonton,Alberta started making noises as described no sooner than they were beginning to plan their own internet service. I hear they have a special tariff discount for fax lines (business or residential) too, or at least used to. Are they talking with two tounges ? bye,Kai

1 0

Re: Bell vs. Internet Providers (fwd)
by Scott Mace 10 Nov '95

10 Nov '95

> > > ISP's should consider using a DS1 or better to deliver dialtone for their > > modem banks. It makes alot of sense for the ISP and the phone company. > > Delivering dialtone over a T1 is certainly more cost effective than paying > > a %400 increase for Centrex lines. > > What does that have to do with anything? The T1 still drops into a phone > switch and uses up channels. T1's are designed to stay up all the time, and centrex lines aren't. It doesn't really solve the capacity problem, but when Bell sells a T1 they sell it with the full knowledge that the line will be using all channels all the time (unless its a frac T). From what I read, the major concern was problems in planning for switch capacity. Bell really doesn't have a way of knowing that the Centrex lines they just sold to a company (that just happens to be an ISP) are going to be up all the time. This would not be the case if the information collected from the Bell sales rep filters all the way to the people doing the planning (which I doubt it does). I can see Bell charging more for a Centrex line, but I can't see Bell charging more for a T1 or T3 _just_ because an ISP is using it. Scott

2 1

Bell vs. Internet Providers (fwd)
by Paul Ferguson 10 Nov '95

10 Nov '95

>From the 'For What Its Worth' Department. - paul BELL VS. INTERNET PROVIDERS Bell Canada plants to increase line rate charges for Internet service providers by as much as 400%, saying the higher charge is necessary because those providers use lines for 55 to 60 minutes every hour compared with voice usage on a Centrex line of about 10 minutes. The service providers say the move aims to eliminate competition for Bell when it introduces its Worldlinx Internet access by the end of this year. (Montreal Gazette 9 Nov 95 D7) ***************************************************************** Edupage, 9 Nov 95. Edupage, a summary of news items on information technology, is provided three times each week as a service by Educom, a Washington, D.C.-based consortium of leading colleges and universities seeking to transform education through the use of information technology. *****************************************************************

7 6