- NANOG - lists.nanog.org

Re: T1 Circuit actual throughput 1290Kbps
by prue＠ISI.EDU 09 Jul '98

09 Jul '98

Ron, >It isn't D4 or ESF framing that causes a loss in usable bandwidth on a >T1, but AMI line coding. As stated above, when using AMI line coding, >the T1 will not be provisioned for clear channel (64k) DS0s, but will be >provisioned for 56k DS0s, thus providing an aggregate bandwidth of 1344 >kbps. The least significant zero of each DS0 is forced to a 'one', thus >insuring correct ones density on the line. The better solution is B8ZS >coding, which allows full use of each DS0's bandwidth. >Inverting the data on the router won't give you back the "lost" 192 kbps >of bandwidth, since you're still limited to using 56kb/s of each DS0. Sorry but you are wrong. AMI has a ones density requirement which causes CSU's to inject ones if the data that is presented it doesn't have enough ones transitions. Running HDLC data inverted can and does cause enough transitions so that the CSU won't inject a one (injecting an error in the data) to keep the phone company multiplexers happy. Several early CSU's had various options to permit you to run at speeds above 1.344 Mb/s including a scrambling method (not fool proof), various proprietary encoding methods, and the hdlc invert option. In all cases if the encoding method failed to provide one's density then the CSU circuitry would inject the required ones. Note that the ones stuffing in every low order channel bit exceeds the ones density requirement of AMI ciruits. Most AMI circuits require no more than 15 zero bits in a row and at least three one bits in every 24 bits. I believe that there was the FCC part 68 rule and the 62411 rule if I remember it right. They were slightly different. Most early CSU manufacturers took the easy implementation method of low order bit ones stuffing, but that was not the only solution nore the most bandwidth efficient solution to assuring ones density and data intetrity. However if you are running on a channelized T1 circuit then you really are limited to 56k X 24 or 1.344 Mb/s. That is not related to D4 or AMI. You would see that with an ESF circuit as well.

1 0

Re: T1 Circuit actual throughput 1290Kbps
by Ron Buchalski 09 Jul '98

09 Jul '98

>Date: Thu, 9 Jul 1998 10:18:51 -0400 (EDT) >From: "Robert E. Seastrom" <rs(a)bifrost.seastrom.com> >To: dorian(a)blackrose.org >CC: ltd(a)interlink.com.au, tonyh(a)noc.cbn.net.id, nanog(a)merit.edu >Subject: Re: T1 Circuit actual throughput 1290Kbps > > > From: Dorian Kim <dorian(a)blackrose.org> > > > A T1 is capable of achieving 1536 kbps maximum (24 x 64 kbps). > > While this doesn't seem to apply to Tony's case, I wouldn't make a blanket > statements like that. If the T1 is provisioned ESF, yes you can get > 1536 kbps, but there are places where you still can only get SF/D4 framing. > >Precisely. The most likely explanation is that the T1 is actually D4 >framed, which gives it a throughput of (24 * 56kbps) == 1344k. Or the >T1 may be properly provisioned but the CSU/DSUs incorrectly configured >for the old-fashioned framing (I think I saw this working at one >point, can't remember for sure). The PPP is hardly eating anything at >all in the grand scheme of things. > > ---Rob It isn't D4 or ESF framing that causes a loss in usable bandwidth on a T1, but AMI line coding. As stated above, when using AMI line coding, the T1 will not be provisioned for clear channel (64k) DS0s, but will be provisioned for 56k DS0s, thus providing an aggregate bandwidth of 1344 kbps. The least significant zero of each DS0 is forced to a 'one', thus insuring correct ones density on the line. The better solution is B8ZS coding, which allows full use of each DS0's bandwidth. Inverting the data on the router won't give you back the "lost" 192 kbps of bandwidth, since you're still limited to using 56kb/s of each DS0. -rb Ron Buchalski SE, Cisco Systems ______________________________________________________ Get Your Private, Free Email at http://www.hotmail.com

1 0

Re: T1 Circuit actual throughput 1290Kbps
by prue＠ISI.EDU 09 Jul '98

09 Jul '98

>In some cases, if you can't get B8ZS line coding, and have to use AMI, >you might have to drop to 24x56=1344kbps (and use one bit on each DS0 >to maintin 1's density), but even then, if your routers or DSUs can >invert the data, you can still go 24x64=1536kbps. (HDLC guarantees >zero-density ... so if you invert it, you get guaranteed 1's density.) Yep. I've done that since 1988 on old circuits that were delivered before B8ZS was available. It does work. However you still can not get 1.536 over a T1 in general, close but not quite. Cisco framing as well as PPP framing uses HDLC encoding. The rules for HDLC encoding say that any time there are 5 one bits in a row within the data portion of a frame, a zero bit is inserted by the transmitter and removed by the receiver. This adds to the overhead of the T1 link. It is of course possible to send data that never has 5 one bits in a row, but real data will have this. The 5 ones don't have to be only within a byte. The contiguous one bits can cross byte boundries to incur this ones stuffing algorithm overhead. Walt

1 0

Cisco scripts??
by Tim Kempka 09 Jul '98

09 Jul '98

Does anyone know where I can get ahold of some sort of script that querys the routers via SNMP that shows the running configs on them? I want to have a web based interface that ONLY shows the running configs. I am aware of the built in http server that is on the routers, but I would rather do it via SNMP . Any information would be appreciated. Thanks, Tim Kempka ____________________________________________________ Tim Kempka E-Mail: tim(a)flash.net Network Engineer Pager: page_tim(a)flash.net FlashNet Communications Phone: 817-589-2390 ext. 378

8 7

Re: T1 Circuit actual throughput 1290Kbps
by Sean Butler 09 Jul '98

09 Jul '98

Tony Hariman wrote: >Does anybody have experience having a T1 circuit with PPP >encapsulation getting only 1290 Kbps maximum throughput looking >at "sh int" result from cisco router or MRTG ? Is it a long delay link running TCP traffic? Long delay links will always show significanlty lower utilization than the max theoretical... You can check out the following paper for more details, but I'll give a quick example below. Lakshman&Madhow "The Performance of TCP/IP for Networks with High BWxDelay products and random loss" ,IEEE Trans. on Networking Vol5(3), june 97, pp329-335 http://tesla.csl.uiuc.edu/~madhow/publications/ton97.ps >From that paper, you can create the following chart: Buffer Size Link Utilization 60K 81% 120K 87% 180K 91% 480K 99% The percentage utilization is based on the bandwidth * the delay, so if you have a T1 at 1536K with a delay of 400ms, you get just over 600K (very pessimistic... -- you can play with the numbers based on the delay of the link. My numbers are based on an analysis I did of a satellite link some time ago...) The buffer sizes are percentages of that 600K, at 10%, 20%, etc. The buffer size is also the queue length * the max packet size of the buffer.... So a Cisco with a default queue length of 40 and an MTU of 1500 = 60K, at which point you'd only see 81% utilization for 1 TCP session... That is just 1 session, so you'd expect to see higher utilization for more sessions. This all assumes FIFO queing... Another point to consider is that with Internet traffic, there are many many 'small flows' where it takes time to ramp up the send window size, and before the time it would ramps up to use the whole pipe, the flow is done Sean Butler

1 0

Re: Cisco scripts??
by Ron Johnson 09 Jul '98

09 Jul '98

I will have to say, that rtrmon is one of the best tools that we have for management of our routers. Simple, elegant, hard to port to non BSD, does all kinds of routine tasks easy. > > Does anyone know where I can get ahold of some sort of script that querys > > the routers via SNMP that shows the running configs on them? I want to > > have a web based interface that ONLY shows the running configs. I am aware > > of the built in http server that is on the routers, but I would rather do > > it via SNMP . Any information would be appreciated. > > I know someone said it can be done with SNMP but we don't do it that way here. > > http://www.vix.com/rtrmon/ > http://www.vix.com/vulture/ > > Both of these were paid for by Genuity but are freely redistributable. > -- > Paul Vixie > La Honda, CA "Many NANOG members have been around > <paul(a)vix.com> longer than most." --Jim Fleming > pacbell!vixie!paul (An H.323 GateKeeper for the IPv8 Network)

3 2

Re: Cisco scripts??
by Paul Vixie 08 Jul '98

08 Jul '98

> Does anyone know where I can get ahold of some sort of script that querys > the routers via SNMP that shows the running configs on them? I want to > have a web based interface that ONLY shows the running configs. I am aware > of the built in http server that is on the routers, but I would rather do > it via SNMP . Any information would be appreciated. I know someone said it can be done with SNMP but we don't do it that way here. http://www.vix.com/rtrmon/ http://www.vix.com/vulture/ Both of these were paid for by Genuity but are freely redistributable. -- Paul Vixie La Honda, CA "Many NANOG members have been around <paul(a)vix.com> longer than most." --Jim Fleming pacbell!vixie!paul (An H.323 GateKeeper for the IPv8 Network)

1 0

Re: Process management
by Sean Donelan 08 Jul '98

08 Jul '98

Ok, I give up, here are the cites (pun intended). The U.S. Military 'lessons learned' web site I refered too is http://redwood.do.langley.af.mil/ The Nuclear Regulatory Commission reports I refered too is http://www.nrc.gov/NRC/GENACT/GC/gcabout.html Information about subscribing to the mailing list is at the bottom of the page. -- Sean Donelan, Data Research Associates, Inc, St. Louis, MO Affiliation given for identification not representation

1 0

Process management
by Sean Donelan 08 Jul '98

08 Jul '98

>Customer Support and process management seems to be the Achilles' heel of >the Internet industry. This industry has seemed to build up a terrible >reputation for poor customer support. Coming from the Nuclear industry, I >am truly jaded when it comes to process management, but the Internet >industry seems to have taken a track as far as possible on the other side. Ok, so I'm wierd. I enjoy reading the incident reports nuclear plant operators have to file. One nuclear power plant operator discovered the diesel fuel for their standby emergency generators was actually syrup. How many ISPs actually test what's in that tanker truck before filling up their generators? Or how about the time all the uninterruptiable power supplies at a plant turned off in unison because of a bug in the UPS firmware. How many ISPs would even know that other people had a problem with a particular brand of UPS, and they should check theirs? Has anyone, anywhere, ever really solved these problems? I don't know, but I suspect not. The nuclear industry generates a LOT of paperwork, but its just a bunch of dead trees if no one reads and learns from it. The nuclear industry seems to suffer from the 'Its not my fault because I followed approved procedures.' The internet industry seems to suffer from the 'Its not my fault because no one controls the Internet.' The airline industry suffers from 'Its always the captin's fault.' The telephone industry suffers from 'Its not my fault, we're the telephone company, we can't be wrong.' And so forth. No one just 'knows' this stuff. The cliches do ring true. You either learn it the hard way from your own mistakes. Or you learn it the easy way from other people's mistakes. I'm not saying mistakes can be eliminated. Too many useful discoveries have resulted from mistakes. AS7007 was pretty nasty, but if we had learned our lesson, it might have been worth it. The problem is we didn't learn our lesson. We really need to work on communicating the lessons learned. The U.S. Military has a wonderful web site on 'lessons learned' which they update almost every day. If you are cynical, you might say that just shows the military makes mistakes almost every day. On the other hand, if the lessons are really learned, it means every day there is one less mistake which will occur in the future. I suspect one of the academic types would love to write a grant proposal to operate a web site for Internet operations lessons learned. The question is would anyone contribute anything to it. Doubtful after the first month. And who gets the book deal with O'Reilly. >There apparently have been several organizations developed recently that >claim to be attempting to address this problem. However, I have not seen >an industry-wide affect by the operations of these organizations. Then let's get together and form one. The standard operating procedure for forming a new organization is everyone puts in $25,000 and the organization puts out a press release announcing the start of a search for an executive director. We can name W.C. Fields as an ex-officio board member so we can say to tiny tykes, "Go away son, you're bothering me." As you point out there are a number of organizations which have popped up. However it seems to be a lot easier to get mega-corporations to contribute funding in exchange for having their name on a 'worthy cause' than it is to do something concrete. With the millions of dollars ISP salespeople have told me they are spending on their customer service center phone switches, why do I continue to have such a problem getting an answer the question "What phone number should I call if I want to report a problem?" -- Sean Donelan, Data Research Associates, Inc, St. Louis, MO Affiliation given for identification not representation

2 1

Re: small vent
by dlr＠bungi.com 08 Jul '98

08 Jul '98

[In the message entitled "Re: small vent" on Jun 28, 14:54, Jon Lewis writes:] > On Sun, 28 Jun 1998, Mark Milhollan wrote: > > > "David R. Conrad" writes: > > >b) someone has dummied up a prefix in a mail header or something (typical) > > >c) there is a bug in the database software (happens) > > > > I doubt B since the IP address queried was that of the TCP peer. > > Granted it could have been a spoofed session, but that would have > > required a level of expertise that I've never seen used merely to > > attempt delivery of a few thousand pieces of e-mail. > > What about d). Perhaps someone temporarily announced routes for > unallocated space, setup a mail/spam server in that IP space, sent out > their mail, and stopped exporting the route. > > Is there a place one can search to see a history of BGP announcements for > a given route? This has, in fact, happened before. One of the reasons that the unallocated spaces are listed on the RBL. -- Dave Rand dlr(a)bungi.com http://www.bungi.com

3 2