Free full BGP feed. IPv4 and IPv6 (2020) | Hacker News Reader

Free full BGP feed. IPv4 and IPv6 (2020)

lukasz.bromirski.net

48 pointspm222237 comments

[shameless plug] I made a tool to look into BGP data as well as other networky things at https://god.ad. Can be added to homescreen as a PWA app.

https://god.ad/view/AS15169

this is great, thank you.

Thanks, looks very llmish

Doesn’t mean it’s bad. Thanks for the work

Front-end was indeed done with Claude, with many iterations and tweaks though

does anyone know where can i buy rDNS records? purpose is find all domains hosted on specific ip for a monthly price? I think crawling all my own is big environment waste of resources, i'd be happy to pay for a subscription or somethign

> does anyone know where can i buy rDNS records? purpose is find all domains hosted on specific ip for a monthly price

Are rdns records always populated? I thought they were optional and many sites didn't bother

i am talking about this data: https://host.io/lgyy.cc

They mean anyone will only ever be able to get some rather than all because of how DNS works. Even that site is unable to find all such mappings for every IP. I.e. the first option to build such a dataset is to use rDNS lookups, but it's optional to configure those and many don't bother (opting to just have the forward lookup). The other way about it is collecting domains and seeing which IPs they resolve to, but this is only ever partial (there is no such thig as a conplete lists of domains in the world) and has additional problems with CDNs or things hosted in more than one place at once (you need to do a lookup from every possible location each domain could conceivably be hosted to be sure you get all of the reverse mappings for that domain).

As an example: Unless you know my domains ahead of time you'll never be able to come up with what domains are hosted on my IPs because I don't bother to configure rDNS. So those IPs will look like they host no servers (or only some if you only had a partial list of my domains) rather than all of them on those IPs.

Anyways, for free data sources trying to get a partial view of this you can check out Rapid7's Sonar or Common Crawl. Each should have the pieces needed to construct this kind of view from the data.

Sonar does seem to have data here: https://opendata.rapid7.com/

But it seems you've to go through their sales team and all for the data.

I can't find CommonCrawl making their DNS resolution available.

Common Crawl doesn't publish raw DNS separately, you have to pull the information out of the aggregate database. The WARC-IP-Address header should record the IP Common Crawl connected to for the site.

Good timing, I'm about to release that dataset.

Do you mean rDNS records or a mapping of an IP address to all ([semi] publicly known) DNS records that point to an IP address? Cause these are two different things. Just because a.com points to 1.2.3.4 does not mean 1.2.3.4 has an rDNS record to a.com.

You are right, i'd need both actually. I can't find any service which provides weekly/monthly data download for a flat subscription price, so that i can perform local fast lookups.

How granular do you need this data? Domains are ultimately not publicly known, especially when it comes to subdomains. Do you just need domains that get a certain amount of public traffic or are you trying to get as much domains as possible? In the latter case scraping this data would require the logs of public DNS servers to even know of the existence of many domains (although certificate transparency logs could get one half the way).

Just to have a compressed static file that contains all the routes one could download from time to time I think would be interesting to analyze. Same goes for the anycast root DNS servers. To have a full dump from them could be interesting. Not to be confused with the root.zone [1] I mean the whole kit and caboodle.

[1] - https://www.internic.net/domain/root.zone

RIPE keeps this https://ris.ripe.net/docs/mrt/#name-and-location

The collectors are browsable nice, Thankyou!

The collectors seem to be overloaded so made a temporary mirror of the one for AMS In AMS. [1] I only mirrored the latest files. Feel free to beat it up. The .gz files decompress to just under 4GB

[1] - https://ams.nochan.net/data.ris.ripe.net/rrc00/

No interest in this mirror, gonna nuke it.

This is conflating BGP routes and DNS.

DNS data: Root server data is available via AXFR ("dig . AXFR @f.root-servers.net") but this isn't what you're referencing.

Second level TLD server data is available is available at CZDS. (https://czds.icann.org/home) but some TLDs don't participate, but this also isn't what you're referencing.

What I think you want: There is no canonical list of all zones that exist - there is no "central repository" once you pass the root zone downwards - that's a feature, not a bug. Some organizations have partial views based on large recursive resolver data (DomainTools, Google, Cisco, Cloudflare, Quad9) but access to that data is limited to vetted researchers or more typically only available at a cost (disclaimer: I work for Quad9.) Smaller versions of recursive data sets exist, but are usually significantly limited by geography and demographics of the user community that generates the data set.

BGP route data: This exists in many forms in realtime like the site referenced above, though historic data is difficult to track. No matter what the source or latency, there is bias in the data set because BGP pathing is unique to each ASN that collects it - no two views of the table are identical, and any data set is as "best guess" at state conditions at that time.

Here are some possible data sets for BGP:

Packet Clearing House (PCH) provides a set of snapshots going back 20+ years (though it seems to be offline at the moment): https://www.pch.net/resources/Routing_Data/

Cymru has a live version you can query via various APIs (including ironically via DNS): https://www.team-cymru.com/ip-asn-mapping

Routeviews from University of Oregon also has a data set that is widely used by researchers: https://www.routeviews.org/routeviews/

This is conflating BGP routes and DNS.

No I was fantasizing that I could get these two things. They are two separate things. If I had a nickle for every time I said something off topic I could by everyone a cup of coffee.

> Just to have a compressed static file that contains all the routes one could download from time to time I think would be interesting to analyze.

When I needed this for some work stuff, it was pretty easy to find table dumps and work with them? I don't remember where it was, but I'm sure you can find some. After an acquisition, we had our own ASN and I was able to get table dumps from our own infrastructure.

> Same goes for the anycast root DNS servers. To have a full dump from them could be interesting. Not to be confused with the root.zone [1] I mean the whole kit and caboodle.

What do you think the root servers have that's not in the root.zone? I think you can AXFR from the root servers, too, but it should be the same thing as from the https site.

What do you think the root servers have that's not in the root.zone?

Every glue record for every domain name. The glue is the valuable goo. Without the goo everyone goes back to /etc/hosts which I am not opposed to.

The root servers don't have the glue record for every domain name though. All they have is the glue records for the TLDs. (which is in the zone file you linked)

Every time I stand up new name servers I have to add the glue records into the root servers or the name servers do not exist. In fairness to me I keep forgetting the trend is to shove everything into big centralized DNS servers as it is something I would never do at least not as a primary.

Afaik, those glue records are held at your TLD's registry and served by the tld nameservers, not the root servers.

It might be nice to get a zone transfer for every tld, but that's not possible for the public. (I understand there's some way to get many of them, but $$$$)

I run my own name servers. I never use the name servers of a registry. I can see the glue records of my name servers in the root servers. In fact the reason I left NetworkSolutions (web.com) was that their interface to update the root servers broke and there was nobody left that knew how to fix it. I'm sure they must have fixed it by now but I was being impatient only waiting 3 weeks.

I should add that I have been adding name servers to the root servers since 1998. I've just never managed one of the root servers and I guess nobody on HN has either.

I'm pretty sure you're confused.

If you query your domain at the root servers, they will refer you to the tld servers, run by your registry.

Then when you query your domain at the tld servers, they return your selected nameservers along with their addresses, if in their bailiwick.

Im not confused at all. For your recursive servers to know how to get the name servers they have to have a hints file, that tells them what to use for the tld, then from the group of root servers associated with that TLD you get the glue (name servers) assocociated with the domain, then you query the authoritative name servers for that domains resources you were requesting. Every name server has a glue record in the root servers. I can manually walk you through it with dig if you would like.

I need to clean up the formatting a bit but this is a walk through. [1] I think I know where your confusion is but I will wait for your reply.

[1] - https://blawg.nochan.net/b/Internet-Crap/20260530-DNS-Recurs...

> from the group of root servers associated with that TLD you get the glue (name servers) assocociated with the domain

Those are not root servers. They are tld servers, run by the registry for the tld your domain is under.

The root servers serve the root. Root hints just tell you where those are. The zone file you listed has the whole root zone.

a.root-servers.net is a root server.

a.gtld-servers.net is a tld server (controlled by verisign, iirc?). several tlds use gtld-servers, but that doesn't make them root servers.

TLD servers are just another level of root servers. Each TLD does have a primary registrar that manages them however they provide access to thousands of registrars to manage the glue records at that layer. They are still collectively part of the root servers even if they are not under the domain root-servers.net. My registrar does not manage them. Nobody here is on a regisrar that manages the TLD root servers. The only permission my registrar and most registrars have is to update glue records for the domains their domain clients own and to add their domains. I and every DNS admin I've ever know collectively refers to all levels of the root and TLD servers as root servers. Even if that is technically incorrect that is how we have always referred to them. I think that distinction would primarily mater for people that worked at one of the high level registrars that manages the anycast clusters of tld root servers. At least I think that is where the confusion started. This was fun we should do this again. I updated the blawg to add (TLD) in reference to root servers.

If every level is a root, then the word root means nothing. I guess you could say that the root servers point to the tld root servers that point to your domain's root servers. But then root server just means dns server.

I never mentioned registrars, only registries.

> Nobody here is on a regisrar that manages the TLD root servers.

I have a .is domain. For .is and some of the other country code tlds, you contract directly with the entity managing the TLD and its servers.

Are you secretly the admin of archive.is? He's made accounts on here before but they were all temporary. I want to believe he has an account that has been here a long time.

No, I live on an island, and is is a usps acceptable abbreviation for island ... I was trying to convince my spouse to leave yahoo mail for mail hosted on one of our domains by getting a domain that's less out there than enslaves.us, so I have islandname.is (well not literally that).

Tldr, a cool domain name isn't enough to make a switch happen.

Seems to be from 6 years ago.

Ignorant question here: Isn't this an open port for BGP poisoning?

Only if they haven't enabled filtering, which would be ludicrous.

Reply on news.ycombinator.com