- Test - lists.nanog.org

Re: MAE-East still no generator
by Peter Lothberg 08 Sep '96

08 Sep '96

> > I realize that the phone world that MFS is part of operates > > differently, but this is networking equipment we are talking about, > > not phone switches, and UPSes are very old tech at this point. > > the bottom layers of the network (L2 and below) are mostly telco > stuff and mostly (other than the gigaswitch) prefer DC power. DC Our Gigaswitches run on DC..... --Peter

1 0

Re: MAE-East still no generator
by Peter Lothberg 08 Sep '96

08 Sep '96

> >|} Check out the nice graph of MAE-East for that last 5 days. When are > >|} they going to get a generator installed? Why can't you pay for a > >|} generator when you are charging +30 people $5,700 a month? > > > >At least part of the facility (1919 Gallows, P1) is on generator, or > >hasn't had a power outage recently. > > > >mae-east uptime is 3 weeks, 3 days, 5 hours, 46 minutes > > All of our equipment is on battery-backed -48 volt DC power. Well, for > full disclosure, almost all of it, there is a dialup modem plugged into > a wall wart; but I can't find an rackmount DC-powered NEBS compliant > modem in quantities of one. > > But, there seems to be some piece of equipment somewhere that's missing > power. Even though the FDDI interface stays operational, we seem to loose > almost all the peering sessions. Even other folks who also are on DC power. > AS286 sticks around for some reason I can't explain. I guess its time to > physically walk the circuit path looking for equipment plugged into a > wall outlet. > > BTW, are other folks getting advanced notice from MFS about this? Our > sales person claims we're on the notification list, but we still don't get > advanced warning about work at mae-east. > -- > Sean Donelan, Data Research Associates, Inc, St. Louis, MO > Affiliation given for identification not representation > ICM-MAE-E uptime is 2 weeks, 4 days, 14 hours, 12 minutes System restarted by reload at 14:40:57 EDT Tue Aug 20 1996 Running buggy software Neighbor V AS MsgRcvd MsgSent TblVer InQ OutQ Up/Down State 192.41.177.1 4 1 92312 159041 6773476 0 0 2d17h 192.41.177.115 4 2548 31408 457020 6773476 0 0 2d17h 192.41.177.140 4 690 116540 212363 6773476 0 0 1d10h 192.41.177.141 4 1673 99241 158848 6773501 0 0 2d17h 192.41.177.170 4 3830 103181 211298 6773501 0 0 2d07h 192.41.177.180 4 3561 221435 210611 6773501 0 0 2d17h 192.41.177.181 4 3561 26754 210272 6773501 0 0 2d17h 198.67.131.89 4 1800 763230 569598 6773631 0 0 2d11h This is funny, I did loose my IBGP peers, maybe the MFS provided IDSU is on AC power... -:) --Peter

1 0

Re: MAE-East still no generator
by Vadim Antonov 08 Sep '96

08 Sep '96

Paul A Vixie <vixie(a)wisdom.home.vix.com> wrote: >DC >power is what the box wants internally, and there's a lot less heat >generated by making 5V and 12V out of 48VDC instead of 110VAC or 220VAC. >if you visit a telco switching center and surround yourself with 10 >Northern Telecom OC48 racks you should notice how quiet it seems. Well, it is not exactly true. The way modern power supplies work is to rectify AC to DC, modulate it with high frequency (1KHz or so) with some transistors to feed a small transformer (the higher frequency, the smaller core you need), and then rectify it again. The power loss on transistors working in saturated modes is mostly due the voltage loss (which is constant determined by physics of the PN-junction or field-effect) multiplied by the current. With lower voltage you need higher current to get the same power, so the power dissipation on switching transistors is higher. Therefore, high-voltage power supplies are more efficient than low-voltage. The reason why sync mux racks are quiet is because they don't do any fancy high-speed stuff (after all, those muxes are nothing more than glorified shift registers) and so they don't need any power-hungry circuitry. --vadim /

1 0

Re: MAE-East still no generator
by Peter Lothberg 08 Sep '96

08 Sep '96

So, can someone explain what the problem is? MAE-East is a telco facility (well...) but if I look at one of our routers; ICM-MAE-E uptime is 2 weeks, 4 days, 14 hours, 12 minutes System restarted by reload at 14:40:57 EDT Tue Aug 20 1996 So, what's the fuzz here? (Don't you all use -48Vdc in telco pops?) --Peter

2 1

Re: MAE-East still no generator
by Sean Donelan 08 Sep '96

08 Sep '96

>I realize that the phone world that MFS is part of operates >differently, but this is networking equipment we are talking about, >not phone switches, and UPSes are very old tech at this point. > >In other words, its silly for them not to have UPSed AC available. I know I'm going to regret stating what I thought was obvious, but when has that ever stopped me. When ordering and installing equipment you need to be aware of the environment you are placing the equipment. When installing equipment in a normal office, use standard 120 AC powered boxes because thats the power source you find in a normal office. When installing equipment in large data centers, the best power comes from three-phase 208-volt outlets with strange plugs. Order the right equipment for the job, or you might get stuck trying to fit a square plug in a round hole. When installing equipment in a telco facility the best power comes from -48 volt DC power sources. Once again, ordering the right equipment for the environment it will be housed will avoid many problems later. Its true telcos don't have much experience running data centers for customers, formally known as timesharing, now known as co-locating. But the folks who do, EDS, IBM, Digital, etc, sense the gap in the market and are moving quickly to fill it. Maybe the trend will reverse and telcos will start co-locating at data centers instead of data centers co-locating at central offices. Operationally, "web hosting" isn't that much different from the timesharing services of twenty years ago. The question for network operators is who do you think will have an easier time. The billion dollar computer companies bringing data centers up to central office communications standards. Or the billion dollar telephone companies bringing central offices up to data center standards. If you think you know the answer, place your bets, err, I mean orders. -- Sean Donelan, Data Research Associates, Inc, St. Louis, MO Affiliation given for identification not representation

1 0

Re: MAE-East still no generator
by matthew＠scruz.net 08 Sep '96

08 Sep '96

Original message <199609080429.XAA10713(a)westie.gi.net> From: Alan Hannan <alan(a)gi.net> Date: Sep 7, 23:28 Subject: Re: MAE-East still no generator > > > Howdy, > > To stay in the tradition of the NANOG mailing list, I will take > this subject a bit off topic. :-) > > The common American household and business operates on Alternating > Current mainly because of distance. > > Over long distances, Alternating Current loses less energy in > transferring energy than Direct Current. For proof, imagine the > energy actually moving from point a -> point b (as in DC) or > moving back and forth in millions of sets between points a and b. > (and not moving as far) (as in AC). The real reason that AC is more efficient to transmit is that transformers work on AC, and that means you can, with little effort, trade amps for volts. Then you get to transmit the power at high voltage, but low current draw. That means thin wires (saves copper) and less resistive loss (saves energy). Are we far enough off topic yet? -matthew kaufman matthew(a)scruz.net

2 1

Re: MAE-East still no generator
by Sean Donelan 08 Sep '96

08 Sep '96

>|} Check out the nice graph of MAE-East for that last 5 days. When are >|} they going to get a generator installed? Why can't you pay for a >|} generator when you are charging +30 people $5,700 a month? > >At least part of the facility (1919 Gallows, P1) is on generator, or >hasn't had a power outage recently. > >mae-east uptime is 3 weeks, 3 days, 5 hours, 46 minutes All of our equipment is on battery-backed -48 volt DC power. Well, for full disclosure, almost all of it, there is a dialup modem plugged into a wall wart; but I can't find an rackmount DC-powered NEBS compliant modem in quantities of one. But, there seems to be some piece of equipment somewhere that's missing power. Even though the FDDI interface stays operational, we seem to loose almost all the peering sessions. Even other folks who also are on DC power. AS286 sticks around for some reason I can't explain. I guess its time to physically walk the circuit path looking for equipment plugged into a wall outlet. BTW, are other folks getting advanced notice from MFS about this? Our sales person claims we're on the notification list, but we still don't get advanced warning about work at mae-east. -- Sean Donelan, Data Research Associates, Inc, St. Louis, MO Affiliation given for identification not representation

1 0

Re: Administrative Bounds
by randy＠psg.com 07 Sep '96

07 Sep '96

> Take a routing dump and look at all of the consecutive /24 with the > eact same AS path. This isn't the registry's fault. The lack of aggregation has been mentioned on NANOG recently, with folk being very quiet. When you take a look at it, it is a mess with some of the classic chest-thumpers being significant offenders. Tony, what happened to the analysis you used to post? We need it again. Badly. randy

2 1

Re: Hey, anyone who wanna help improve the net stability?
by Justin W. Newton 07 Sep '96

07 Sep '96

At 03:01 PM 9/5/96 -0500, Alan Hannan wrote: >] Of the class cs from the swamp how many are now being routed at the >] defaultless core? HOW MANY ADDITIONAL CLASS Cs FROM THE SWAMP ARE THERE >] FOR WHICH THE OWNERS COULD DEMAND ROUTING? > > Lots. Uhm, I dunno that people can really /demand/ routing. Ask for it, expect it or pray for it yes, but there is no reason that ISP X won't come up with a policy tomorrow saying we have a policy saying that we won't route any addresses that come from Justin. Justin Newton Internet Architect Erol's Internet Services

2 1

Re: customers and web servers and level one naps
by dgaudet＠hotwired.com 07 Sep '96

07 Sep '96

In article <hot.mailing-lists.nanog-199609051951.PAA08990(a)netaxs.com>, Avi Freedman <freedman(a)netaxs.com> wrote: >The bigger problem (the one I was thinking about) would be someone putting >a host on 192.41.177.x or 192.157.69.x or 198.32.136.x or ... With something like cisco's Local Director you could quite easily serve a whole web farm out of a single IP. Well, at least when HTTP/1.1 style Host: or full-url GET type requests become more prevalent, otherwise you'd be stuck with a single virtual host. Technically it's a router :) Dean

1 0