As the author states the data is coming from the peering DB datasets, are you implying there is an error in the extractions/the datasets themselves? or that the data in the peering db datasets does not correlate to the real world? I believe the later but wanted to make sure I am not missing something

Example : https://burdantes.github.io/cdn_facility_evol_timeseries.html Double click on Yahoo to select that trace. It states that Yahoo only exists in 2 interconnection facilities since December of 2023. Now take a look at PeeringDB : https://www.peeringdb.com/net/27 . We're in a lot more than 2, and have been for at least a decade, maybe two. My assumption is that the author likely used one of our OTHER ASNs in PeeringDB , not AS10310 which is our primary one. If *my* data displayed is so wildly incorrect, it's fair to question how accurate the data may or may not be for anyone else. On Mon, Mar 31, 2025 at 10:39 AM Brian Turnbow <brian.turnbow@cdlan.it> wrote:

Hi Tom,

Il giorno lun 31 mar 2025 alle ore 15:53 Tom Beecher via NANOG <nanog@lists.nanog.org> ha scritto:

...

Some of the data presented by the author is.... um... just wrong. Take

it

...

with a very large grain of salt.

As the author states the data is coming from the peering DB datasets, are you implying there is an error in the extractions/the datasets themselves? or that the data in the peering db datasets does not correlate to the real world? I believe the later but wanted to make sure I am not missing something

Thanks

Brian

Hi Tom, I sincerely apologize for this mistake. We used an existing dataset that associated ASNs with providers, along with some manual checks and additions. We validated the results for the providers we were initially focused on, but unfortunately, Yahoo wasn’t part of that initial validation. When we later added it, we didn’t go through the same verification process. So, while our dataset included a set of ASNs for Yahoo, it was missing the key one you pointed out, and we didn’t catch the error. We’ll be implementing additional checks across all providers to improve our accuracy, but we can’t guarantee that every ASN will be captured. We’ll also update our figures and are happy to incorporate any fixes that others flag. Thanks again for bringing this to our attention! Best, Loqman

Appreciate the apology, but that's not necessary. Just a friendly suggestion, the natural reaction from internet nerds when we see something presented with significant errors is to discount most (if not all) of the presented work. ( "If they missed that, what else did they miss?") It's worth extra caution to be as diligent as possible to avoid that. Also appreciate the fix in the graphs. Wasn't really needed; I of course know what we have, and as much as it pains me to say, I don't think Yahoo really belongs in any current conversations about 'hypergiants'. We were once, but not really in that category today. Either way , good luck and take care! On Mon, Mar 31, 2025 at 9:43 PM Loqman Salamatian via NANOG < nanog@lists.nanog.org> wrote:

I wanted to follow up to confirm that we’ve updated the files with the corrected mapping (see https://burdantes.github.io/cdn_facility_evol_timeseries.html), and we’re also working on updating them in the blog post. If you spot any other mistakes, please don’t hesitate to reach out! _______________________________________________ NANOG mailing list

https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/JDXSVWHN...