- NANOG - lists.nanog.org

RE: Network Availability Report
by Jason Lixfeld 01 Dec '00

01 Dec '00

Depends on how well it works, as far as I'm concerned. -----Original Message----- From: Kevin_Badeen(a)harte-hanks.com To: Jason Lixfeld Cc: nanog(a)nanog.org; owner-nanog(a)merit.edu Sent: 01/12/00 11:06 AM Subject: Re: Network Availability Report How much momney do you want to spend? Kevin Badeen Unix Administrator Harte-Hanks Response Management |--------+-----------------------> | | Jason Lixfeld| | | <jlixfeld@tea| | | m.look.ca> | | | Sent by: | | | owner-nanog@m| | | erit.edu | | | | | | | | | 12/01/2000 | | | 09:59 AM | | | | |--------+-----------------------> >----------------------------------------------------------------------- ----------------------------------------------------------------------| | | | To: "'nanog(a)nanog.org'" <nanog(a)nanog.org> | | cc: | | Subject: Network Availability Report | >----------------------------------------------------------------------- ----------------------------------------------------------------------| I'm looking for some kind of tool that can generate a report on network availability. For instance; this utility would run on a server and perform tests to other hosts and routers across the network. The usual Network Monitoring Tool tests like ping, port response, etc but also things like throughput between monitoring host and destination, packet-loss and a total percentage of availability for that host. Obviously the availability "number" would be dynamic based on the results of the tests and how the user configures the impact of those tests against the available percentage. Has anyone seen anything like this in one nice little package? What I'm going to do is give the head honchos a URL where all they have to do is click on a portion of our network and read off the uptime percentage. Thx! ------------------------------------------------------------------------ - This email server is running an evaluation copy of the MailShield anti- spam software. Please contact your email administrator if you have any questions about this message. MailShield product info: www.mailshield.com

1 0

Re: Network Availability Report
by Kevin_Badeen＠harte-hanks.com 01 Dec '00

01 Dec '00

How much momney do you want to spend? Kevin Badeen Unix Administrator Harte-Hanks Response Management |--------+-----------------------> | | Jason Lixfeld| | | <jlixfeld@tea| | | m.look.ca> | | | Sent by: | | | owner-nanog@m| | | erit.edu | | | | | | | | | 12/01/2000 | | | 09:59 AM | | | | |--------+-----------------------> >---------------------------------------------------------------------------------------------------------------------------------------------| | | | To: "'nanog(a)nanog.org'" <nanog(a)nanog.org> | | cc: | | Subject: Network Availability Report | >---------------------------------------------------------------------------------------------------------------------------------------------| I'm looking for some kind of tool that can generate a report on network availability. For instance; this utility would run on a server and perform tests to other hosts and routers across the network. The usual Network Monitoring Tool tests like ping, port response, etc but also things like throughput between monitoring host and destination, packet-loss and a total percentage of availability for that host. Obviously the availability "number" would be dynamic based on the results of the tests and how the user configures the impact of those tests against the available percentage. Has anyone seen anything like this in one nice little package? What I'm going to do is give the head honchos a URL where all they have to do is click on a portion of our network and read off the uptime percentage. Thx!

1 0

GPO web site - supremecourtus.gov
by Sean Donelan 01 Dec '00

01 Dec '00

It appears the Government Printing Office is in the eye of the storm again on Friday. The GPO hosts the US Supreme Court web site, and if you remember GPO also distributed the Starr report. Its will be interesting to see how much the GPO has learned, and how well they handle the interest in the website on Friday. Good Luck Genuity.

2 1

Internet metrics
by Schmiedt, Jamie 01 Dec '00

01 Dec '00

I am looking for websites that deal with bandwidth and performance metrics as they pertain to the Internet? Anything and everything goes. I need to get a snapshot in time of things like "the average speed of the Internet", the average size of e-mail messages, the average size of a web page, peering points and hosting facilities, current bottlenecks be they hardware, software or networks. Any cool links for this sorta info will be greatly appreciated. Thanks in advance -jamie

2 1

Re: Traceroute versus other performance measurement
by smd＠clock.org 30 Nov '00

30 Nov '00

Mmm, network analysis stuff on NANOG. Scary! | Do I understand it correctly that the "invariant 5 tuple" refers to | 5 non-variable aspects of this performance test, being namely: Tuple definition, from the CS Jargon file, a copy of which is here: http://rkb.home.cern.ch/rkb/AN16pp/node289.html Note that a 5-tuple is basically a set of 5 elements/objects/records. The "invariant 5-tuple" here just has to do with 5 things which don't change in the test. There is another element which is unchanged during a ttcp run: TOS/diffserve, thus there is really a (or an at-least) 6-tuple invariance. In practice, only 5 of the elements in that n-tuple are used in making a forwarding decision; the choice of the fifth could well be vendor/release-dependent, at least judging by what people are saying here. | E.g.: performance measurements will gravitate towards a result not | representative for the totality of all traffic, as the 5 variables | are rather fixed, or at the least fixed for the duration of the test ? Well, if you are making a big flow, and the elements used by routers to make a hash which is used to select which of several equal-cost paths to use don't change, then the routers will use the same single path throughout the ttcp test. (Even if the hash->path mapping is adjusted, the ttcp test packets hash the same, so path flutter doesn't really change the principle). Thus if you have six 10Mbps lines in parallel, the ttcp packets will hash the same way, and use a single 10Mbps link (at a time). This is good, because TCP does not like out-of-order packets, and because traffic loads are fractal, out-of-order packets are likely where there are multiple paths, and if there are many multiple-paths this can lead to the TCP sender losing its brains with duplicate ACKs pushing the congestion window right through the floor. | Has intelligent switching and forwarding (Layer-4 and up) started to | make such a significant impact on end-to-end performance that all | traditional performance tests must necessarily be flawed ? Research question. My guess: probably not. A better question, maybe the one you want to ask, is how well does the Internet deal with lots of equal-cost paths? The answer: small numbers are OK, large numbers are probably not OK, and moreover, the goal of avoiding out-of-order TCP packets constrains any given TCP flow to one path, which means that that TCP flow will not be able to run any faster than the slowest component, across any given link. If that component is a bottleneck, oops. | (there must be a reason why my MRTG graphs of HTTP GET's to websites | are so much more consistent than icmp pings...) Consistent? I'm not following. Maybe there is a large-numbers issue here: if you are plotting lots and lots and lots and lots of flows, and only looking at a few pings, then if there is frequent very-short-term queueing delays at work, you will see them more clearly with the individual pings. A large number of pings aggregated together will probably converge upon whatever you see when you aggregate a large number of HTTP GETs. cf. http://skepdic.com/lawofnumbers.html and the cool toy at http://www.stat.berkeley.edu/~stark/Java/lln.htm Note that the tendency of small numbers of short-term samples to be wildly different than large-number samples when looking at ping (which ultimately is a measurement of queueing behaviour) is an important property of the Internet. The short-term behaviour is _fractal_, and is fractal across a surprising number of _generations_ (where a generation is, for example, 0.001 sec to 0.01 sec, that is an order of magnitude increase base ten). This really really messes up the fine-tuning of problem identification, because again the law of large numbers suggests that as the Internet grows in terms of numbers of users, more of them will see shockingly bad behaviour if they ping somewhere for a few seconds. :-( (In other words, Danie Senie was right with respect to users seeing strange things and asking about them). Sean.

2 1

Re: Will the lights stay on - Predictions about the US electrical (fwd)
by multics＠ruserved.com 30 Nov '00

30 Nov '00

At the time I worked on the contract with NYSPA around 1976, the reliability goal was one failure in 10 years. This did not include failures from accidents, like cars knocking down power poles, lightning strikes, etc. >From what I've seen at least where I am in NYS, we've had 2 2 hour outages in the last 6 years, one unscheduled, one scheduled for maintenance. The phone company hasn't been as reliable. Forwarded message: > On Tue, 28 November 2000, Bill Woodcock wrote: > > > Congress within the next five years will enact legislation mandating > > > minimum standards for reducing downtime and improving system > > > reliability > > > > Jeez, if that's all it takes, why didn't they do it years ago? :-) > > > NEW YORK, Nov 29 (Reuters) - New Jersey, in an effort to encourage utilities to improve the reliability of the state's electric system, adopted, late Tuesday, standards designed to reduce outage frequency and duration. > > http://biz.yahoo.com/rf/001129/n29106533.html > > > Essentially every book on security, quality control, reliability, etc includes > a statement "First, you must have the attention of the top management." How > do you get the attention of the top management? Will passing a law or a > regulation improve reliablity? It could, if it got the attention of the > management of the companies to increase the budget for doing the things > better reliability requires. > > I'm at the damage prevention convention this week. All the major carriers > had people on one of the panels on the problem of fiber cuts. The scariest > thing is most carriers only have 1 or 2 people in their entire organization > responsible for damage prevention. Worldcom has 2 people watching 60,000 > router miles. I think Level 3 had 1 person. Qwest International had 2 people. > ATT has a department, but even that seems to be downsized. When carriers are > merged, the damage prevention budget is usually cut in half. > > I was very impressed with the people from all the carriers. They are all > working very hard. But how much can 1 or 2 people really do? Fundamentally, > if you want to improve reliability you must make reliability important to > the top management of a company. Make it worth the budget management must > approve to hire the people to implement the reliability program. > > Do you know what I want in an SLA? > > I would prefer my circuit never fails, but when it does I don't want money. I > want a public apology from the CEO of the company. CEO's at major companies > tend to have very strong personalities. They hate to ever admit any problems. > Notice when carriers issue press releases with good news, it always > includes a quote from the CEO. But when they release bad news, it is always > a spokesperson. Maybe if the CEO had to deal with the fall out of reliability > problems, he or she would decide increasing the budget a bit is worth it. > > After Ebay's CEO had to appear on CNBC, suddenly reliability became important > for Ebay. After AOL's CEO had to appear, suddenly reliability became a > top management concern of AOL and they spent a lot of money on it. > > > > -- Richard Shetron multics(a)ruserved.com multics(a)acm.rpi.edu NO UCE What is the Meaning of Life? There is no meaning, It's just a consequence of complex carbon based chemistry; don't worry about it The Super 76, "Free Aspirin and Tender Sympathy", Las Vegas Strip.

1 0

Re: Will the lights stay on - Predictions about the US electrical grid
by Sean Donelan 30 Nov '00

30 Nov '00

On Tue, 28 November 2000, Bill Woodcock wrote: > > Congress within the next five years will enact legislation mandating > > minimum standards for reducing downtime and improving system > > reliability > > Jeez, if that's all it takes, why didn't they do it years ago? :-) NEW YORK, Nov 29 (Reuters) - New Jersey, in an effort to encourage utilities to improve the reliability of the state's electric system, adopted, late Tuesday, standards designed to reduce outage frequency and duration. http://biz.yahoo.com/rf/001129/n29106533.html Essentially every book on security, quality control, reliability, etc includes a statement "First, you must have the attention of the top management." How do you get the attention of the top management? Will passing a law or a regulation improve reliablity? It could, if it got the attention of the management of the companies to increase the budget for doing the things better reliability requires. I'm at the damage prevention convention this week. All the major carriers had people on one of the panels on the problem of fiber cuts. The scariest thing is most carriers only have 1 or 2 people in their entire organization responsible for damage prevention. Worldcom has 2 people watching 60,000 router miles. I think Level 3 had 1 person. Qwest International had 2 people. ATT has a department, but even that seems to be downsized. When carriers are merged, the damage prevention budget is usually cut in half. I was very impressed with the people from all the carriers. They are all working very hard. But how much can 1 or 2 people really do? Fundamentally, if you want to improve reliability you must make reliability important to the top management of a company. Make it worth the budget management must approve to hire the people to implement the reliability program. Do you know what I want in an SLA? I would prefer my circuit never fails, but when it does I don't want money. I want a public apology from the CEO of the company. CEO's at major companies tend to have very strong personalities. They hate to ever admit any problems. Notice when carriers issue press releases with good news, it always includes a quote from the CEO. But when they release bad news, it is always a spokesperson. Maybe if the CEO had to deal with the fall out of reliability problems, he or she would decide increasing the budget a bit is worth it. After Ebay's CEO had to appear on CNBC, suddenly reliability became important for Ebay. After AOL's CEO had to appear, suddenly reliability became a top management concern of AOL and they spent a lot of money on it.

4 3

Traceroute versus other performance measurement
by Paul Bradford 30 Nov '00

30 Nov '00

I have been reading NANOG posts for probably 2 years now.. this is my 1st post. I need help with a reality/sanity check. Traceroute is a good tool for checking for routing type problems (loops). Does anyone feel it's a good tool to use for testing "bandwidth".... My obvious answer is it isn't a good tool for that.... One problem I see is that the way traceroute works, if a transport mixes media between say Ethernet to LANE and back to Ethernet you give room for Destination unreachable responses from a trace route because you have to to packet switching medias with a fast cell switched media in between.... packets less than 64k (like traceroute info) are easily lost in the conversion from ethernet to LANE. Does this sound right? Thanks, Paul A. Bradford

8 7

STM4c and Cisco
by Ariel Biener 30 Nov '00

30 Nov '00

Hi guys/gals, For a certain operation, we will be ordering a STM4c circuit. The edge router (our side) is a Cisco 7507, running RSP8, with VIP4-80s. I saw that Cisco havea PA-OC12 card for the 750x series. My questions is as follows: This particular location is an Internet exchange. Traffic can be classified as general purpose, and I assume web access is the largest share of it. I heard of different limitations of the VIP4-80, as per how many pps it can do, and what my boss heard from Cisco was about 400Mbit/s worth of packets (if we take the usual packet size distribution into consideration). I wonder if this is true. AFAIK, there are PA-GE cards for 750x series too. Would they also be stuck at ~40% of their potential ? For that OC12 card, this makes it roughly 500Mb/s worth of cells (tax this for packet to cell conv. and you're at 400Mbit/s). Does any of you run STM4c circuits on 750x routers (not 12k or Junipers) ? If you do, have you ever came close to the full capacity of the link ? I need to know this before I order it (although this is more academic than anything else, since I have no choice in the matter anyway). thanks for you input, --Ariel -- Ariel Biener e-mail: ariel(a)post.tau.ac.il PGP(6.5.8) public key http://www.tau.ac.il/~ariel/pgp.html

2 1

MACs on PAIX
by Matt Hempel 30 Nov '00

30 Nov '00

I'm using MAC accounting on our Fast E connection at the PAIX to do per-peer traffic stats. It worked great until we put in a GSR. I don't know why it worked before or if anything has changed since. The stats aren't working now because the # of entries in the MAC accounting table exceed 512 (a hard IOS limit). The majority of these aren't in the ARP table ... they are inbound without translation ... just layer 2 noise. It seems odd to me that there are this many distinct addresses on the same shared segment. Not to mention that the port isn't switched. To wit: 0000.0000.0000(0 ): 110444951 packets, 27399M bytes, last: 80ms ago 0000.0003.0000(3 ): 9416 packets, 2738919 bytes, last: 455680ms ago 0000.037f.4129(20 ): 33 packets, 7700 bytes, last: 302129568ms ago Pretty serious chunks of bandwidth. There are a large group of these addresses that begin with 0000. --- are these some sort of bridged translation? Can you think of any other reason that so many MACs would appear on the same port? --matt hempel

3 2