- mat-wg - mailman.ripe.net

Format change for index pages of RIS MRT data files and RIPE Atlas daily dataset dumps
by Oleg Muravskiy 24 Nov '22

24 Nov '22

Dear RIS and RIPE Atlas users, On Wednesday, 30 November 2022 we are planning to migrate the web server at <https://data.ris.ripe.net/> as well as <https://data-store.ripe.net> from Apache httpd to nginx. This will affect how the file listings in HTML index pages are generated, including the following: * RIS MRT data files (updates and dumps) * RIPE Atlas daily dataset dumps If you rely on the format of these pages, you need to update your software. Before we do the upgrade, you could check the new format of the index pages at: https://test-data.ris.ripe.net/rrc00/ We would like to ask you to avoid relying on the specific format of the index.html files. If you are fetching BGP dump files, please use the REST API that is described at: https://ris.ripe.net/docs/40_Prototypes/15_per_rrc_dumps.html You might also consider using per-peer dump files instead: https://ris.ripe.net/docs/40_Prototypes/10_per_peer_dumps.html We want to encourage you to provide your feedback on the API, so that we could improve it and build a similar API for the update files. If you are fetching update files, please use direct URLs to files as described at: https://ris.ripe.net/docs/20_raw_data_mrt.html#name-and-location Kind regards, Oleg Muravskiy, Robert Kisteleki RIPE NCC

1 0

streaming link MAT session ripe85
by Massimo Candela 27 Oct '22

27 Oct '22

The MAT session started at RIPE85. Streaming here: https://ripe85.ripe.net/live/side/

1 0

MAT agenda for RIPE85
by Massimo Candela 26 Oct '22

26 Oct '22

Dear all, I'm pleased to announce that we have an agenda for RIPE85! https://ripe85.ripe.net/programme/meeting-plan/mat-wg/ Ciao, Massimo (on behalf of the MAT chairs)

2 1

Re: [mat-wg] [ippm] [Rpm] [M-Lab-Discuss] misery metrics & consequences
by J Ignacio Alvarez-Hamelin 26 Oct '22

26 Oct '22

Dear all, After some time in silence on the IPPM list, I like to make some comments here. As we presented in the draft-ietf-ippm-route-00 (now the RFC9198), the main problem is the traffic follows heavy-tailed distributions when it is seen from the end-to-end points: the origin of most of the issues in that video. Therefore, treating it as parametric distribution is not possible, unless you are dealing with a complex distribution like the Stable distribution: • B. Mandelbrot, “New methods in statistical economics,” Journal of political economy, vol. 71, no. 5, pp. 421–440, 1963. • ——, “The variation of certain speculative prices,” The journal of business, vol. 36, no. 4, pp. 394–419, 1963. (and so it will be extremely complex a high computing demands.) This is why we propose to use quartiles to characterize delays in the RFC9198. Then, I am doing some research to understand how the delay can change with network load, using the quartiles. I attach some measurements done during the pandemic, showing the congestion as a function of the time. Best, Ignacio _______________________________________________________________ Dr. Ing. José Ignacio Alvarez-Hamelin CONICET and Facultad de Ingeniería, Universidad de Buenos Aires Av. Paseo Colón 850 - C1063ACV - Buenos Aires - Argentina +54 (11) 5285 0716 / 5285 0705 e-mail: ihameli(a)cnet.fi.uba.ar web: http://cnet.fi.uba.ar/ignacio.alvarez-hamelin/ _______________________________________________________________ > On 23 Oct 2022, at 08:57, Sebastian Moeller <moeller0(a)gmx.de> wrote: > > Hi Glenn, > > >> On Oct 23, 2022, at 02:17, Glenn Fishbine via Rpm <rpm(a)lists.bufferbloat.net> wrote: >> >> As a classic died in the wool empiricist, granted that you can identify "misery" factors, given a population of 1,000 users, how do you propose deriving a misery index for that population? >> >> We can measure download, upload, ping, jitter pretty much without user intervention. For the measurements you hypothesize, how you you automatically extract those indecies without subjective user contamination. >> >> I.e. my download speed sucks. Measure the download speed. >> >> My isp doesn't fix my problem. Measure what? How? >> >> Human survey technology is 70+ years old and it still has problems figuring out how to correlate opinion with fact. >> >> Without an objective measurement scheme that doesn't require human interaction, the misery index is a cool hypothesis with no way to link to actual data. What objective measurements can be made? Answer that and the index becomes useful. Otherwise it's just consumer whining. >> >> Not trying to be combative here, in fact I like the concept you support, but I'm hard pressed to see how the concept can lead to data, and the data lead to policy proposals. > > [SM] So it seems that outside of seemingly simple to test throughput numbers*, the next most important quality number (or the most important depending on subjective ranking) is how does latency change under "load". Absolute latency is also important albeit static high latency can be worked around within limits so the change under load seems more relevant. > All of flent's RRUL test, apple's networkQuality/RPM, and iperf2's bounceback test offer methods to asses latency change under load**, as do waveforms bufferbloat tests and even to a degree Ookla's speedtest.net. IMHO something like latency increase under load or apple's responsiveness measure RPM (basically the inverse of the latency under load calculated on a per minute basis, so it scales in the typical higher numbers are better way, unlike raw latency under load numbers where smaller is better). > IMHO what networkQuality is missing ATM is to measure and report the unloaded RPM as well as the loaded the first gives a measure over the static latency the second over how well things keep working if capacity gets tight. They report the base RTT which can be converted to RPM. As an example: > > macbook:~ user$ networkQuality -v > ==== SUMMARY ==== > Upload capacity: 24.341 Mbps > Download capacity: 91.951 Mbps > Upload flows: 20 > Download flows: 16 > Responsiveness: High (2123 RPM) > Base RTT: 16 > Start: 10/23/22, 13:44:39 > End: 10/23/22, 13:44:53 > OS Version: Version 12.6 (Build 21G115) > > Here RPM 2133 corresponds to 60000/2123 = 28.26 ms latency under load, while the Base RTT of 16ms corresponds to 60000/16 = 3750 RPM, son on this link load reduces the responsiveness by 3750-2123 = 1627 RPM a reduction by 100-100*2123/3750 = 43.4%, and that is with competent AQM and scheduling on the router. > > Without competent AQM/shaping I get: > ==== SUMMARY ==== > Upload capacity: 15.101 Mbps > Download capacity: 97.664 Mbps > Upload flows: 20 > Download flows: 12 > Responsiveness: Medium (427 RPM) > Base RTT: 16 > Start: 10/23/22, 13:51:50 > End: 10/23/22, 13:52:06 > OS Version: Version 12.6 (Build 21G115) > latency under load: 60000/427 = 140.52 ms > base RPM: 60000/16 = 3750 RPM > reduction RPM: 100-100*427/3750 = 88.6% > > > I understand apple's desire to have a single reported number with a single qualifier medium/high/... because in the end a link is only reliably usable if responsiveness under load stays acceptable, but with two numbers it is easier to see what one's ISP could do to help. (I guess some ISPs might already be unhappy with the single number, so this needs some diplomacy/tact) > > Regards > Sebastian > > > > *) Seemingly as quite some ISPs operate their own speedtest servers in their network and ignore customers not reaching the contracted rates into speedtest-servers located in different ASs. As the product is called internet access I a inclined to expect that my ISP maintains sufficient peering/transit capacity to reach the next tier of AS at my contracted rate (the EU legislative seems to agree, see EU directive 2015/2120). > > **) Most do by creating load themselves and measuring throughput at the same time, bounceback IIUC will focus on the latency measurement and leave the load generation optional (so offers a mode to measure responsiveness of a live network with minimal measurement traffic). @Bob, please correct me if this is wrong. > > >> >> >> On Fri, Oct 21, 2022, 5:20 PM Dave Taht <dave.taht(a)gmail.com> wrote: >> One of the best talks I've ever seen on how to measure customer >> satisfaction properly just went up after the P99 Conference. >> >> It's called Misery Metrics. >> >> After going through a deep dive as to why and how we think and act on >> percentiles, bins, and other statistical methods as to how we use the >> web and internet are *so wrong* (well worth watching and thinking >> about if you are relying on or creating network metrics today), it >> then points to the real metrics that matter to users and the ultimate >> success of an internet business: Timeouts, retries, misses, failed >> queries, angry phone calls, abandoned shopping carts and loss of >> engagement. >> >> https://www.p99conf.io/session/misery-metrics-consequences/ >> >> The ending advice was - don't aim to make a specific percentile >> acceptable, aim for an acceptable % of misery. >> >> I enjoyed the p99 conference more than any conference I've attended in years. >> >> -- >> This song goes out to all the folk that thought Stadia would work: >> https://www.linkedin.com/posts/dtaht_the-mushroom-song-activity-69813666656… >> Dave Täht CEO, TekLibre, LLC >> >> -- >> You received this message because you are subscribed to the Google Groups "discuss" group. >> To unsubscribe from this group and stop receiving emails from it, send an email to discuss+unsubscribe(a)measurementlab.net. >> To view this discussion on the web visit https://groups.google.com/a/measurementlab.net/d/msgid/discuss/CAA93jw4w27a…. >> _______________________________________________ >> Rpm mailing list >> Rpm(a)lists.bufferbloat.net >> https://lists.bufferbloat.net/listinfo/rpm > > _______________________________________________ > ippm mailing list > ippm(a)ietf.org > https://www.ietf.org/mailman/listinfo/ippm

2 1

Re: [mat-wg] [Rpm] [ippm] [M-Lab-Discuss] misery metrics & consequences
by J Ignacio Alvarez-Hamelin 25 Oct '22

25 Oct '22

Dear Bob, I answer inline. Thank you for your comments. P.D. It seems that I have no rights to send emails to the Measurement Analysis and Tools Working Group: how I can join it? _______________________________________________________________ Dr. Ing. José Ignacio Alvarez-Hamelin CONICET and Facultad de Ingeniería, Universidad de Buenos Aires Av. Paseo Colón 850 - C1063ACV - Buenos Aires - Argentina +54 (11) 5285 0716 / 5285 0705 e-mail: ihameli(a)cnet.fi.uba.ar web: http://cnet.fi.uba.ar/ignacio.alvarez-hamelin/ _______________________________________________________________ > On 25 Oct 2022, at 14:02, rjmcmahon <rjmcmahon(a)rjmcmahon.com> wrote: > > I don't understand the information in the link. It looks like lines on a map to some form of dials which are too small to read. I suppose that depends on the zoom of your browser, try to put 100% (command+ to zoom-in in chrome). You will see a 24 hour clock where some hours are green (no congestion), yellow (some congestion) or red (congestion). But all of that is just an example. > > One can sample and create a Gaussian per the central limit theorem (CLT) if the underlying process probability density functions converge, i.e. can be integrated to 1. With that said, normalizing does lose information and doesn't say much about tails, outliers, etc. The Stable distribution is a generalization over all the distributions. Consider that Gaussian has finite moments, and heavy-tailed distribution not, this is where Stable one enters as a generalization. > > We should be careful in assuming only the tails matter and that all traffic follows heavy-tailed distributions. With bufferbloat it's the minimum of the latency PDF that shifts. Codel watches a minimum. "Jacobson suggested that average queue length actually contains no information at all about packet demand or network load.[3][5] He suggested that a better metric might be the minimum queue length during a sliding time window." Well, it depends on which point of the network you really are, but there a lot where this is true. In the other hand, no heavy tailed distributions do not cause problems if you use statistical tools that suport anything. For example, the median (Q2) converges to the mean (fist statistical moment) for Poisson distributions. > > We need statistical tools that also allow for the analysis of non-parametric distributions too. Hotelling T2 assumes the multivariate distributions are Gaussian. Kolmogorov-Smirnov tests can be used for non-parametric distributions. We find both are needed for SPC used by our automation systems. Yes, I did a lot of this kind of test, and I found very interesting things, even in the very important links (important ISP with important CDN), verifying Stable distributions with certain parameters (even some ones with heavy-tailed distributions). The problem is how to do this a low computable cost, and this why you can use quartiles. > > Sample subgroups of one really don't give sufficient information about any type of distribution, parametric or non-parametric. Here I cannot understand: which are the subgroups? > > Bob >> Dear all, >> After some time in silence on the IPPM list, I like to make some >> comments here. As we presented in the draft-ietf-ippm-route-00 (now >> the RFC9198), the main problem is the traffic follows heavy-tailed >> distributions when it is seen from the end-to-end points: the origin >> of most of the issues in that video. Therefore, treating it as >> parametric distribution is not possible, unless you are dealing with a >> complex distribution like the Stable distribution: >> • B. Mandelbrot, “New methods in statistical economics,” Journal of >> political economy, vol. 71, no. 5, pp. 421–440, 1963. >> • ——, “The variation of certain speculative prices,” The journal of >> business, vol. 36, no. 4, pp. 394–419, 1963. >> (and so it will be extremely complex a high computing demands.) >> This is why we propose to use quartiles to characterize delays in the >> RFC9198. Then, I am doing some research to understand how the delay >> can change with network load, using the quartiles. >> You can see some measurements done during the pandemic, showing the >> congestion as a function of the time (24 hours maximum): >> https://cnet.fi.uba.ar/ignacio.alvarez-hamelin/RIPE-Atlas-measurement-24681… >> [you can zoom in and out, pan it, clicking on the Xs you can close >> dialogs, to reopen them click on the link] >> Best, >> Ignacio >> ___________________________________ >> _______________________________________________________________ >> Dr. Ing. José Ignacio Alvarez-Hamelin >> CONICET and Facultad de Ingeniería, Universidad de Buenos Aires >> Av. Paseo Colón 850 - C1063ACV - Buenos Aires - Argentina >> +54 (11) 5285 0716 / 5285 0705 >> e-mail: ihameli(a)cnet.fi.uba.ar >> web: http://cnet.fi.uba.ar/ignacio.alvarez-hamelin/ >> _______________________________________________________________ >>> On 23 Oct 2022, at 08:57, Sebastian Moeller <moeller0(a)gmx.de> wrote: >>> Hi Glenn, >>>> On Oct 23, 2022, at 02:17, Glenn Fishbine via Rpm <rpm(a)lists.bufferbloat.net> wrote: >>>> As a classic died in the wool empiricist, granted that you can identify "misery" factors, given a population of 1,000 users, how do you propose deriving a misery index for that population? >>>> We can measure download, upload, ping, jitter pretty much without user intervention. For the measurements you hypothesize, how you you automatically extract those indecies without subjective user contamination. >>>> I.e. my download speed sucks. Measure the download speed. >>>> My isp doesn't fix my problem. Measure what? How? >>>> Human survey technology is 70+ years old and it still has problems figuring out how to correlate opinion with fact. >>>> Without an objective measurement scheme that doesn't require human interaction, the misery index is a cool hypothesis with no way to link to actual data. What objective measurements can be made? Answer that and the index becomes useful. Otherwise it's just consumer whining. >>>> Not trying to be combative here, in fact I like the concept you support, but I'm hard pressed to see how the concept can lead to data, and the data lead to policy proposals. >>> [SM] So it seems that outside of seemingly simple to test throughput numbers*, the next most important quality number (or the most important depending on subjective ranking) is how does latency change under "load". Absolute latency is also important albeit static high latency can be worked around within limits so the change under load seems more relevant. >>> All of flent's RRUL test, apple's networkQuality/RPM, and iperf2's bounceback test offer methods to asses latency change under load**, as do waveforms bufferbloat tests and even to a degree Ookla's speedtest.net. IMHO something like latency increase under load or apple's responsiveness measure RPM (basically the inverse of the latency under load calculated on a per minute basis, so it scales in the typical higher numbers are better way, unlike raw latency under load numbers where smaller is better). >>> IMHO what networkQuality is missing ATM is to measure and report the unloaded RPM as well as the loaded the first gives a measure over the static latency the second over how well things keep working if capacity gets tight. They report the base RTT which can be converted to RPM. As an example: >>> macbook:~ user$ networkQuality -v >>> ==== SUMMARY ==== >>> Upload capacity: 24.341 Mbps >>> Download capacity: 91.951 Mbps >>> Upload flows: 20 >>> Download flows: 16 >>> Responsiveness: High (2123 RPM) >>> Base RTT: 16 >>> Start: 10/23/22, 13:44:39 >>> End: 10/23/22, 13:44:53 >>> OS Version: Version 12.6 (Build 21G115) >>> Here RPM 2133 corresponds to 60000/2123 = 28.26 ms latency under load, while the Base RTT of 16ms corresponds to 60000/16 = 3750 RPM, son on this link load reduces the responsiveness by 3750-2123 = 1627 RPM a reduction by 100-100*2123/3750 = 43.4%, and that is with competent AQM and scheduling on the router. >>> Without competent AQM/shaping I get: >>> ==== SUMMARY ==== >>> Upload capacity: 15.101 Mbps >>> Download capacity: 97.664 Mbps >>> Upload flows: 20 >>> Download flows: 12 >>> Responsiveness: Medium (427 RPM) >>> Base RTT: 16 >>> Start: 10/23/22, 13:51:50 >>> End: 10/23/22, 13:52:06 >>> OS Version: Version 12.6 (Build 21G115) >>> latency under load: 60000/427 = 140.52 ms >>> base RPM: 60000/16 = 3750 RPM >>> reduction RPM: 100-100*427/3750 = 88.6% >>> I understand apple's desire to have a single reported number with a single qualifier medium/high/... because in the end a link is only reliably usable if responsiveness under load stays acceptable, but with two numbers it is easier to see what one's ISP could do to help. (I guess some ISPs might already be unhappy with the single number, so this needs some diplomacy/tact) >>> Regards >>> Sebastian >>> *) Seemingly as quite some ISPs operate their own speedtest servers in their network and ignore customers not reaching the contracted rates into speedtest-servers located in different ASs. As the product is called internet access I a inclined to expect that my ISP maintains sufficient peering/transit capacity to reach the next tier of AS at my contracted rate (the EU legislative seems to agree, see EU directive 2015/2120). >>> **) Most do by creating load themselves and measuring throughput at the same time, bounceback IIUC will focus on the latency measurement and leave the load generation optional (so offers a mode to measure responsiveness of a live network with minimal measurement traffic). @Bob, please correct me if this is wrong. >>>> On Fri, Oct 21, 2022, 5:20 PM Dave Taht <dave.taht(a)gmail.com> wrote: >>>> One of the best talks I've ever seen on how to measure customer >>>> satisfaction properly just went up after the P99 Conference. >>>> It's called Misery Metrics. >>>> After going through a deep dive as to why and how we think and act on >>>> percentiles, bins, and other statistical methods as to how we use the >>>> web and internet are *so wrong* (well worth watching and thinking >>>> about if you are relying on or creating network metrics today), it >>>> then points to the real metrics that matter to users and the ultimate >>>> success of an internet business: Timeouts, retries, misses, failed >>>> queries, angry phone calls, abandoned shopping carts and loss of >>>> engagement. >>>> https://www.p99conf.io/session/misery-metrics-consequences/ >>>> The ending advice was - don't aim to make a specific percentile >>>> acceptable, aim for an acceptable % of misery. >>>> I enjoyed the p99 conference more than any conference I've attended in years. >>>> -- >>>> This song goes out to all the folk that thought Stadia would work: >>>> https://www.linkedin.com/posts/dtaht_the-mushroom-song-activity-69813666656… >>>> Dave Täht CEO, TekLibre, LLC >>>> -- >>>> You received this message because you are subscribed to the Google Groups "discuss" group. >>>> To unsubscribe from this group and stop receiving emails from it, send an email to discuss+unsubscribe(a)measurementlab.net. >>>> To view this discussion on the web visit https://groups.google.com/a/measurementlab.net/d/msgid/discuss/CAA93jw4w27a…. >>>> _______________________________________________ >>>> Rpm mailing list >>>> Rpm(a)lists.bufferbloat.net >>>> https://lists.bufferbloat.net/listinfo/rpm >>> _______________________________________________ >>> ippm mailing list >>> ippm(a)ietf.org >>> https://www.ietf.org/mailman/listinfo/ippm >> _______________________________________________ >> Rpm mailing list >> Rpm(a)lists.bufferbloat.net >> https://lists.bufferbloat.net/listinfo/rpm

1 0

Re: [mat-wg] [ippm] [Rpm] [M-Lab-Discuss] misery metrics & consequences
by J Ignacio Alvarez-Hamelin 25 Oct '22

25 Oct '22

Dear all, After some time in silence on the IPPM list, I like to make some comments here. As we presented in the draft-ietf-ippm-route-00 (now the RFC9198), the main problem is the traffic follows heavy-tailed distributions when it is seen from the end-to-end points: the origin of most of the issues in that video. Therefore, treating it as parametric distribution is not possible, unless you are dealing with a complex distribution like the Stable distribution: • B. Mandelbrot, “New methods in statistical economics,” Journal of political economy, vol. 71, no. 5, pp. 421–440, 1963. • ——, “The variation of certain speculative prices,” The journal of business, vol. 36, no. 4, pp. 394–419, 1963. (and so it will be extremely complex a high computing demands.) This is why we propose to use quartiles to characterize delays in the RFC9198. Then, I am doing some research to understand how the delay can change with network load, using the quartiles. You can see some measurements done during the pandemic, showing the congestion as a function of the time (24 hours maximum): https://cnet.fi.uba.ar/ignacio.alvarez-hamelin/RIPE-Atlas-measurement-24681… [you can zoom in and out, pan it, clicking on the Xs you can close dialogs, to reopen them click on the link] Best, Ignacio ___________________________________ _______________________________________________________________ Dr. Ing. José Ignacio Alvarez-Hamelin CONICET and Facultad de Ingeniería, Universidad de Buenos Aires Av. Paseo Colón 850 - C1063ACV - Buenos Aires - Argentina +54 (11) 5285 0716 / 5285 0705 e-mail: ihameli(a)cnet.fi.uba.ar web: http://cnet.fi.uba.ar/ignacio.alvarez-hamelin/ _______________________________________________________________ > On 23 Oct 2022, at 08:57, Sebastian Moeller <moeller0(a)gmx.de> wrote: > > Hi Glenn, > > >> On Oct 23, 2022, at 02:17, Glenn Fishbine via Rpm <rpm(a)lists.bufferbloat.net> wrote: >> >> As a classic died in the wool empiricist, granted that you can identify "misery" factors, given a population of 1,000 users, how do you propose deriving a misery index for that population? >> >> We can measure download, upload, ping, jitter pretty much without user intervention. For the measurements you hypothesize, how you you automatically extract those indecies without subjective user contamination. >> >> I.e. my download speed sucks. Measure the download speed. >> >> My isp doesn't fix my problem. Measure what? How? >> >> Human survey technology is 70+ years old and it still has problems figuring out how to correlate opinion with fact. >> >> Without an objective measurement scheme that doesn't require human interaction, the misery index is a cool hypothesis with no way to link to actual data. What objective measurements can be made? Answer that and the index becomes useful. Otherwise it's just consumer whining. >> >> Not trying to be combative here, in fact I like the concept you support, but I'm hard pressed to see how the concept can lead to data, and the data lead to policy proposals. > > [SM] So it seems that outside of seemingly simple to test throughput numbers*, the next most important quality number (or the most important depending on subjective ranking) is how does latency change under "load". Absolute latency is also important albeit static high latency can be worked around within limits so the change under load seems more relevant. > All of flent's RRUL test, apple's networkQuality/RPM, and iperf2's bounceback test offer methods to asses latency change under load**, as do waveforms bufferbloat tests and even to a degree Ookla's speedtest.net. IMHO something like latency increase under load or apple's responsiveness measure RPM (basically the inverse of the latency under load calculated on a per minute basis, so it scales in the typical higher numbers are better way, unlike raw latency under load numbers where smaller is better). > IMHO what networkQuality is missing ATM is to measure and report the unloaded RPM as well as the loaded the first gives a measure over the static latency the second over how well things keep working if capacity gets tight. They report the base RTT which can be converted to RPM. As an example: > > macbook:~ user$ networkQuality -v > ==== SUMMARY ==== > Upload capacity: 24.341 Mbps > Download capacity: 91.951 Mbps > Upload flows: 20 > Download flows: 16 > Responsiveness: High (2123 RPM) > Base RTT: 16 > Start: 10/23/22, 13:44:39 > End: 10/23/22, 13:44:53 > OS Version: Version 12.6 (Build 21G115) > > Here RPM 2133 corresponds to 60000/2123 = 28.26 ms latency under load, while the Base RTT of 16ms corresponds to 60000/16 = 3750 RPM, son on this link load reduces the responsiveness by 3750-2123 = 1627 RPM a reduction by 100-100*2123/3750 = 43.4%, and that is with competent AQM and scheduling on the router. > > Without competent AQM/shaping I get: > ==== SUMMARY ==== > Upload capacity: 15.101 Mbps > Download capacity: 97.664 Mbps > Upload flows: 20 > Download flows: 12 > Responsiveness: Medium (427 RPM) > Base RTT: 16 > Start: 10/23/22, 13:51:50 > End: 10/23/22, 13:52:06 > OS Version: Version 12.6 (Build 21G115) > latency under load: 60000/427 = 140.52 ms > base RPM: 60000/16 = 3750 RPM > reduction RPM: 100-100*427/3750 = 88.6% > > > I understand apple's desire to have a single reported number with a single qualifier medium/high/... because in the end a link is only reliably usable if responsiveness under load stays acceptable, but with two numbers it is easier to see what one's ISP could do to help. (I guess some ISPs might already be unhappy with the single number, so this needs some diplomacy/tact) > > Regards > Sebastian > > > > *) Seemingly as quite some ISPs operate their own speedtest servers in their network and ignore customers not reaching the contracted rates into speedtest-servers located in different ASs. As the product is called internet access I a inclined to expect that my ISP maintains sufficient peering/transit capacity to reach the next tier of AS at my contracted rate (the EU legislative seems to agree, see EU directive 2015/2120). > > **) Most do by creating load themselves and measuring throughput at the same time, bounceback IIUC will focus on the latency measurement and leave the load generation optional (so offers a mode to measure responsiveness of a live network with minimal measurement traffic). @Bob, please correct me if this is wrong. > > >> >> >> On Fri, Oct 21, 2022, 5:20 PM Dave Taht <dave.taht(a)gmail.com> wrote: >> One of the best talks I've ever seen on how to measure customer >> satisfaction properly just went up after the P99 Conference. >> >> It's called Misery Metrics. >> >> After going through a deep dive as to why and how we think and act on >> percentiles, bins, and other statistical methods as to how we use the >> web and internet are *so wrong* (well worth watching and thinking >> about if you are relying on or creating network metrics today), it >> then points to the real metrics that matter to users and the ultimate >> success of an internet business: Timeouts, retries, misses, failed >> queries, angry phone calls, abandoned shopping carts and loss of >> engagement. >> >> https://www.p99conf.io/session/misery-metrics-consequences/ >> >> The ending advice was - don't aim to make a specific percentile >> acceptable, aim for an acceptable % of misery. >> >> I enjoyed the p99 conference more than any conference I've attended in years. >> >> -- >> This song goes out to all the folk that thought Stadia would work: >> https://www.linkedin.com/posts/dtaht_the-mushroom-song-activity-69813666656… >> Dave Täht CEO, TekLibre, LLC >> >> -- >> You received this message because you are subscribed to the Google Groups "discuss" group. >> To unsubscribe from this group and stop receiving emails from it, send an email to discuss+unsubscribe(a)measurementlab.net. >> To view this discussion on the web visit https://groups.google.com/a/measurementlab.net/d/msgid/discuss/CAA93jw4w27a…. >> _______________________________________________ >> Rpm mailing list >> Rpm(a)lists.bufferbloat.net >> https://lists.bufferbloat.net/listinfo/rpm > > _______________________________________________ > ippm mailing list > ippm(a)ietf.org > https://www.ietf.org/mailman/listinfo/ippm

1 0

misery metrics & consequences
by Dave Taht 22 Oct '22

22 Oct '22

One of the best talks I've ever seen on how to measure customer satisfaction properly just went up after the P99 Conference. It's called Misery Metrics. After going through a deep dive as to why and how we think and act on percentiles, bins, and other statistical methods as to how we use the web and internet are *so wrong* (well worth watching and thinking about if you are relying on or creating network metrics today), it then points to the real metrics that matter to users and the ultimate success of an internet business: Timeouts, retries, misses, failed queries, angry phone calls, abandoned shopping carts and loss of engagement. https://www.p99conf.io/session/misery-metrics-consequences/ The ending advice was - don't aim to make a specific percentile acceptable, aim for an acceptable % of misery. I enjoyed the p99 conference more than any conference I've attended in years. -- This song goes out to all the folk that thought Stadia would work: https://www.linkedin.com/posts/dtaht_the-mushroom-song-activity-69813666656… Dave Täht CEO, TekLibre, LLC

2 2

RIPE Atlas, RIPEstat and RIS Quarterly Planning Q4 2022
by Robert Kisteleki 27 Sep '22

27 Sep '22

Dear colleagues, We’ve now published our quarterly planning for RIPE Atlas, RIPEstat and RIS. These are available at: https://www.ripe.net/support/documentation/quarterly-planning The goal of our quarterly planning is to communicate the work we’re doing for the coming months, get your input and start a discussion on other developments we should work on. If you have any input on our plans or want to let us know what you would like to see us work on, please let us know. Kind regards, Robert Kisteleki Research and Development Manager RIPE NCC

1 0

Open House: Exploring RIPEstat
by Christian Teuschel 21 Sep '22

21 Sep '22

Dear colleagues, A quick note to remind you that the RIPE NCC Open House: Exploring RIPEstat is taking place on Wednesday, 28 September at 10:00 (UTC). During this online event, we will talk about what makes this tool a favourite for network operators, CERTs and a wide group of Internet stakeholders. I will be joined by guests who will speak in detail about RIPEstat datasets, a number of new features and the new UI. Check out the details and register for free at: https://www.ripe.net/participate/meetings/open-house/ripe-ncc-open-house-ex… Best regards, Christian Teuschel

1 0

Internet Measurement Day - Tomorrow, 13 September, Online
by Gergana Petrova 12 Sep '22

12 Sep '22

Dear colleagues, A quick note to remind you that the Internet Measurement Day Serbia is taking place tomorrow, 13 September, online. During this online event we will talk about tools on how to measure the Internet, such as RIPE Atlas, the Identifier Technology Health Indicators (ITHI Metrics), Routing Information System (RIS), Resource Public Key Infrastructure (RPKI) and the RIPE NCC’s open data platform RIPEstat. You will also hear from a number of local and international organisations, such as RATEL, SOX, CAIDA and more. Even though the particular focus is on Serbia, a lot of the presentations would be general. Perhaps even the more local presentations would be relevant for those of you who are doing business in neighbouring countries. Check out the agenda and if interested, register for free here: https://www.ripe.net/participate/meetings/internet-measurement-day/internet… Best regards, Gergana -- Gergana Petrova Community Development Manager RIPE NCC

1 0

Fwd: Join Virtual En-ROADS Climate Workshop on September 8!
by Vesna Manojlovic 07 Sep '22

07 Sep '22

Dear MAT-WG, here's an invitation to en_ROADS workshop: https://www.climateinteractive.org/get-involved/webinars/ This is a great showcase for the visualisation of complex & open data, measured over long time periods, analysed and fit into models that are predicting outcomes in the future. I find it highly relevant for the operations of Internet infrastructure, since it has elements of: - water availability https://www.datacenterdynamics.com/en/opinions/why-data-centers-need-to-tal… - sea-level rise https://labs.ripe.net/author/ggm/surviving-the-flood-how-data-centres-and-t… - energy consumption https://www.consultancy.uk/news/28772/10-ways-businesses-can-reduce-their-d… I hope you find it useful, regards, Vesna

1 0

samknows reports 853 ms of latency under load for US DSL.
by Dave Taht 26 Aug '22

26 Aug '22

US DSL, according to their measurements, has 853 ms of latency under load, as reported to the FCC. pretty good podcast and blog at: https://samknows.com/blog/bufferbloat - where they also discuss their current and planned methods of measurement. Also, speedtest.net is finally reporting on "working latency", and their new speedtest apps showing it to the user. https://www.ookla.com/articles/introducing-loaded-latency It would be really great if y'all could join in on helping us finally solve the bufferbloat problem internet wide. -- FQ World Domination pending: https://blog.cerowrt.org/post/state_of_fq_codel/ Dave Täht CEO, TekLibre, LLC

1 0

New on RIPE Labs: 240/4 As Seen by RIPE Atlas
by Qasim Lone 23 Aug '22

23 Aug '22

Dear colleagues, In this new article on RIPE Labs, we use data from RIPE Atlas probes to investigate the usage of 240/4, a block of IPv4 addresses 'reserved for future use' and formally known as Class E in the wild. https://labs.ripe.net/author/qasim-lone/2404-as-seen-by-ripe-atlas/ Kind regards, Qasim

1 0

Call for presentations MAT WG at RIPE85
by Massimo Candela 15 Aug '22

15 Aug '22

Dear MAT WG, We should start getting ready for RIPE85! This year we have a 1.5 hours slot on Thursday 27 October at 09:00 CEST (draft, things may change). Please, submit your presentation proposal by September 9th by sending an email at mat-wg-chairs(a)ripe.net. We prefer talks of 15min. Include in your email a title, a draft of the presentation, and a paragraph describing what you will talk about and why the topic is of interest for the community. Please, let us know if you are planning to present in person or remotely. Thanks! Cheers, Nina, Stephen, Massimo

1 1

Announcement of new RIS prototype services
by Oleg Muravskiy 05 Aug '22

05 Aug '22

Dear RIS users, Over the past few months, we have created several prototype services that expose RIS data in different ways. These services provide a REST API, and the data they expose is closer to real-time, as compared to our traditional "bview" and "updates" files. These services are: - Per-peer dumps https://ris.ripe.net/docs/40_Prototypes/10_per_peer_dumps.html - Per-RRC dumps https://ris.ripe.net/docs/40_Prototypes/15_per_rrc_dumps.html - Peer metadata https://ris.ripe.net/docs/40_Prototypes/20_peer_metadata.html Please take a look at the documentation (links provided above), try out these new services, and tell us what we can improve! We would love to hear your feedback. Since we're posting this message across multiple mailing lists, we ask you to please share your feedback on the ris-users(a)ripe.net mailing list, or write to us directly at ris(a)ripe.net rather than on this list. Kind regards, Oleg Muravskiy RIPE NCC

1 0

New on RIPE Labs: Measuring Encrypted-DNS Censorship Using OONI Probe
by Anastasiya Pak 04 Aug '22

04 Aug '22

Dear colleagues, The Open Observatory Network Interference (OONI) is a non-profit free software project that aims to empower decentralised efforts in documenting Internet censorship around the world. In collaboration with different partners, Simone Basso and colleagues publish research reports and research papers documenting cases of Internet censorship around the world. In this new article on RIPE Labs, Simone talks through the preliminary results of an encrypted-DNS censorship study currently being conducted by the OONI. Read now on Labs: https://labs.ripe.net/author/bassosimone/measuring-encrypted-dns-censorship… Best regards, Anastasiya Pak Marketing & Communications Officer RIPE NCC

1 0

New on RIPE Labs: Dealing with the Undercurrent of Unwanted Traffic
by Alun Davies 06 Jul '22

06 Jul '22

Dear colleagues, Leslie Daigle and her colleagues over at the Global Cyber Alliance have been collecting data from hundreds of sensors to get a clearer picture of how much unwanted traffic there is out there. How much counts as too much? Read more on RIPE Labs: https://labs.ripe.net/author/leslie-daigle/dealing-with-the-undercurrent-of… Best regards, Alun Davies RIPE Labs Editor RIPE NCC

3 3

RIPE Atlas, RIPEstat and RIS Quarterly Planning Q3 2022
by Robert Kisteleki 30 Jun '22

30 Jun '22

Dear colleagues, We’ve now published our quarterly for RIPE Atlas, RIPEstat and RIS. These are available at: https://www.ripe.net/support/documentation/quarterly-planning We hope that by providing these plans, we can better inform the community of the ongoing work on the services we operate and get feedback on our work. If you have any input on our plans or want to let us know what you would like to see us work on, please let us know. Kind regards, Robert Kisteleki Research and Development Manager RIPE NCC

1 0

Request: Provide a simple API to get user's network info.
by Darvari 28 Jun '22

28 Jun '22

Hello, One of the main problems in tools that intend to measure censorship or want to measure the quality and security of the user's network is access to a simple API that has several important information of the user's network: - AS Namber - AS Name - Country Code - and if possible, City Name Also the user's IP to be completely erased from the data. To do this, I use an unofficial Cloudflare service that may change or be restricted at any time: https://speed.cloudflare.com/meta <https://speed.cloudflare.com/meta> I am aware that many other tools and projects need this. Because either their backend server is blocked or other services such as ipinfo, ipconfig, avast, etc, are not available in all countries due to sanctions. Therefore, I ask you to provide such a service because you have the most accessibility in the world. Best Regards, Hossein Darvari

1 0

RIPE Atlas, RIPEstat and RIS Quarterly Planning Q2 2022
by Robert Kisteleki 17 Jun '22

17 Jun '22

Dear colleagues, We are now publishing our quarterly plans for RIPE Atlas, RIPEstat and RIS. These are available at: https://www.ripe.net/support/documentation/quarterly-planning The goals of publishing our plans are to be transparent about the work we will be carrying out in the coming months, to get your input on those plans, and to hear what you would like to see us work on going forward. We will provide status updates as the work progresses and add input from the community to the pages. We will also archive previous plans so you can track how we progress with our work. If you have any input on our plans or want to let us know what you would like to see us work on, please let us know. Kind regards, Robert Kisteleki Research and Development Manager RIPE NCC

2 1

Join our Open House on the RIPE NCC Internet Country Report – Bulgaria, Moldova and Romania
by Alastair Strachan 14 Jun '22

14 Jun '22

Dear colleagues, We are organising an Open House on the Internet Country Report we recently published on Bulgaria, Moldova and Romania. The event will be on Wednesday, 15 June from 10:00-11:30 UTC. Note that the Open House is not limited to people from these three countries but is open to anyone interested in learning about how the region’s Internet is developing. You can find the country report at: https://labs.ripe.net/documents/292/RIPE_NCC_Internet_Country_Report_Bulgar… <https://labs.ripe.net/documents/292/RIPE_NCC_Internet_Country_Report_Bulgar…> Register for free at: https://www.ripe.net/participate/meetings/open-house/ripe-ncc-open-house-in… <https://www.ripe.net/participate/meetings/open-house/ripe-ncc-open-house-in…> I hope to see many of you at there! Best regards, Alastair Strachan RIPE NCC Communtiy Development Officer

1 0

RIPE IPmap v1.1.0: reverse-dns engine
by Chris Amin 18 May '22

18 May '22

Dear colleagues, In November of last year, we released version 1.0.0 of RIPE IPmap. The main feature of that version was the new latency engine, but we also included a prototype of the “reverse-dns” engine, which used very basic pattern matching of hostnames from PTR to provide location guesses. We are pleased to announce that with v1.1.0 the reverse-dns engine has received a major upgrade, allowing it to provide much better geolocation answers through pattern analysis of the entire IPmap “latency” and “worlds” datasets, with daily analysis performed mostly thanks to the sc_hoiho tool from CAIDA’s scamper toolkit. More information about the new engine, as well as links to information on the sc_hoiho tool and its theoretical underpinnings can be found at: https://ipmap.ripe.net/docs/01.manual/#reverse-dns As well as the new and improved reverse-dns engine, the suggestion system has been improved to better combine results from the various geolocation engines. As always, you can get started with the API with the IPmap docs at: https://ipmap.ripe.net/docs/ Note: this mail was sent to ipmap(a)ripe.net yesterday, but I'm aware that not many people have subscribed to that list (yet). If you're interested in IPmap you can sign up to that list at: https://lists.ripe.net/mailman/listinfo/ipmap Kind regards, Chris Amin RIPE NCC

1 0

MAT-WG Agenda RIPE 84
by Brian Trammell (RIPE) 10 May '22

10 May '22

Greetings, all, Here’s the agenda for MAT-WG at RIPE 84, 12:00-13:00 Wednesday 18 May 2022 in Berlin: * Welcome and Agenda Bashing * Data, Data Everywhere, And Now We Stop to Think (Leslie Daigle, 15 min) * Anomaly Alerting for RIPE Atlas (Wolfram Friele, 10 min) * Flow Monitoring in bwNet (Daniel Nägele, 10 min) * Tools Update (Robert Kisteleki, 15 min) * Co-Chair Welcome and Farewell (MAT-WG chairs, 10 min) Cheers, Brian, for the RIPE MAT-WG co-chairs

1 0

[TMA 2022 PhD School] Network Intelligence and Measurements - SIGCOMM Travel Grants Available
by Giovane C. M. Moura 20 Apr '22

20 Apr '22

TL;DR: * PhD school in Internet Measurements in Network Intelligence * SIGCOMM travel grants available! (please send to your students) * Application deadline: April 22nd, 2022 * School dates: June 27th and 28th, 2022 * Location Enschede, NL. * https://tma.ifip.org/2022/phd-school/ ===== 10th TMA PhD School Network Intelligence and Measurements June 27th and 28th, 2022 Enschede, The Netherlands https://tma.ifip.org/2022/phd-school/ After the interruption of the last two editions of the TMA PhD school due to the COVID pandemic, we are extremely happy to announce the 10th TMA PhD School, organized by the University of Twente in Enschede, The Netherlands, together with the 6th Network Traffic Measurement and Analysis Conference. Under the theme Network Intelligence and Measurements, the 10th TMA PhD School will include 4 tutorials, by 5 distinct lecturers, over a 2 days program. Lectures are accompanied by related practical lab sessions, where students have the opportunity to put into practice learned notions. Thanks to SIGCOMM, we will offer a significant number of travel grants, see the details below. Registration and Deadlines Registrations are now open, you can apply at https://hotcrp.info.ucl.ac.be/tma2022phd/. As the PhD school is popular, we encourage students to register early. Our goal is to admit as many students as possible but we do note that we have limited room availability. Application deadline: April 22nd at 23:59 UTC Tentative decision notification: April 29th PhD School Speakers The list of speakers for the 10th TMA PhD school includes: Anna Brunstorm is a Full Professor and Research Manager for the Distributed Systems and Communications Research Group at Karlstad University. She will familiarize the students with experimentation and measurements methodologies for next generation mobile networks, based on her vast experience on applied research. She will also give hands-on training on testbeds and analytics approaches, stemming from her numerous contributions in projects such as H2020 MONROE, NEAT or 5GENESIS. Emile Aben is a System architect at the RIPE NCC, Netherlands. The lecture will provide an overview of the RIPE Atlas Internet data collection system along with a review of the addressed use-cases. In the related lab students will be asked to design and lunch experiments on the Atlas measurement platform, and to process the collected data to address some given problem. Thymen Wabeke and Thijs van den Hout are research engineers with SIDN Labs (.nl registry). They research how machine learning (ML) applications can increase the security of the internet and the Domain Name System (DNS). The lecture will focus on operationalizing a ML model. There are several challenges involved in this process, such as efficiently collecting ground truth labels, analysts needing to understand model inferences, and monitoring model performance so you know when it fails and needs updating. The lecture will include a hands-on training, so that participants can try out some techniques immediately. Pedro Casas is a Senior Scientist at the Austrian Institute of Technology, where he acts as Research Lead in AI4NETS (AI/ML for Networking). He will present a broad overview on the application of AI/ML to network data analysis problems, including network security, anomaly detection, and Quality of Experience (QoE) monitoring. The lecture provides an introduction to the basics of AI/ML, going from more traditional applications to newer paradigms, including deep learning, graph-based learning, generative models, explainable AI, learning in adversarial scenarios, and more. The ultimate goal of the lecture is to motivate – for the newcomers, and to strengthen – for those already in the field, the research in AI4NETS. The PhD School program will include 4 poster sessions where students can present their own (ongoing) PhD work to other students and to the lecturers, and receive individual feedback from them. Poster sessions will take place during breaks. The program will also include an informal interactive session between students and the speakers, with the explicit goal of fostering “vertical” interaction between students and the speakers, and “horizontal” interaction between the students themselves. Students will have the opportunity to interview the speakers, asking for their expert opinion on general aspects (non-technical) of the current research dynamics in the field of traffic monitoring and analysis (e.g., how they see the prospective development of a particular research topic in the near future, or what are the key ingredients that drive the success of an experimental measurement platform, etc). The PhD school program will include social activities on both days, where students, lecturers and attendees of the main TMA conference will have the opportunity to interact informally. A social dinner would be organized on the second day. How to apply and Travel Grants The price for registration for the TMA 2022 PhD school will be fixed soon and will be below 250 Euro. This fee will be levied on all attending students, also those receiving travel grants, as the travel grant rules from our sponsors require us to ask for a small own investment. In order to apply, please submit the following as a single PDF document: the title and abstract of your poster (in the fields required by the submissions system) a current CV with contact information and e-mail address a short personal statement, including: i) a description of the research subject followed by the student, and (ii) information that the applicant feels is relevant to support his/her application, e.g., why the PhD School attendance is important to the applicant’s research and career development; In case you also want to apply for optional travel support, we ask you to also include: a letter from your advisor which should: (i) confirm the applicant’s good standing in the institution; (ii) explain why the applicant would benefit from attending the TMA PhD School the estimated expenses for attending the TMA PhD School (total, and breakdown by travel and lodging), and an indication of which cost items would need to be covered by the grant; if the applicant will be unable to attend the school without a travel grant, they should explain why this is the case (in this case the advisor’s letter should additionally explain the current funding status of the applicant and why the applicant is in need of the travel grant) The organization will prioritize awarding of travel grants to students who will be unable to attend without a grant, and additionally based on diversity and the region from which the applicant has to travel, to ensure representation from traditionally underrepresented regions. In special cases, where the registration fee is also prohibitive, the organization may decide to waive the fee at its discretion. Any decision made by the organization regarding attendance and awarding of travel grants is final. We note that reimbursements of travel grants may take up to several months to complete. Important Notes Please pay attention to the following details about the student travel grant application and reimbursement process: The School will admit a limited number of students. Students enrolled in PhD study programs will receive preference both in the admission to the School and the travel grant awarding. Students admitted to the school are encouraged to attend the TMA conference. We expect a student travel grant will significantly offset air fare and shared hotel accommodation. However, it may not fully cover these expenses, since our desire is to maximize participation by students. In order to obtain reimbursement, you must provide original receipts for air fare and hotel expenses, along with boarding passes (plane or train) for both outbound and inbound journeys. All students participating in the School are required to present a poster describing their research and to actively engage with all activities. Failing to do so will lead to admission rejection. Please provide complete contact information. You will be notified via email, so make sure your email address is correct on your application. Hope to see you all in Enschede for a fantastic TMA PhD School!

1 0

New on RIPE Labs: Detecting DNS Root Manipulation
by Alun Davies 19 Apr '22

19 Apr '22

Dear colleagues, In 2021, when reports emerged that hosts in Mexico were unable to reach whatsapp.net, it was determined that middleboxes were to blame. These devices were intercepting the queries to the root instance hosted in China and sending a bogus reply. Taking this as their starting point, Qasim Lone and colleagues have been investigating the prevalence of middleboxes using RIPE Atlas: https://labs.ripe.net/author/qasim-lone/detecting-dns-root-manipulation/ Best regards, Alun Davies RIPE Labs Editor RIPE NCC

1 0