Data-center traffic: what an ASN is and why it is a signal

An autonomous system in RFC 1930 (March 1996) is a connected group of one or more IP prefixes. One or more operators run it, and it has one clearly defined routing policy. The system has a globally unique number, the ASN. The number takes part in the exchange of external routing information and names the system itself. How to read an address and a provider in a report is covered in how to check traffic by IP and provider. An ASN in this RFC is a unit of routing policy, not a data-center label and not a person’s identifier.

What the number does not promise

The document’s recommended external protocol is BGP. The classic formula from BGP-4, which the RFC retells: a set of routers under one technical administration that looks from the outside like one coherent reachability plan. The RFC does not split systems into “residential” and “hosting”. Seeing an ASN and stamping it “data center” or “apartment” does not follow from this text.

Why data-center traffic comes up next to invalid traffic at all: the MRC addendum (final, June 2020) calls GIVT the traffic recognized routinely, by lists or by standard parameter checks. The examples include known data-center traffic that consistently produces invalid activity but not the route artifacts of live users. Beside it in the same examples are bots, spiders, and other crawlers, checks on transaction parameters, a non-browser User-Agent, prefetch without a live person opening the ad, placement sizes 0×0 and 1×1, and sessions that according to client data cannot render an image, including a headless browser.

The list is for an MRC accreditation audit. It is not mixed with the TAG taxonomy. A GIVT share in Direct is not equated with this list. “Known” in the example means traffic that consistently produces invalid activity and does not look like live users’ route. Not every hosting address is declared invalid by that sentence.

A shared address is not an ASN

RFC 6598 (April 2012) describes a different gluing of addresses. Providers are assigned the space 100.64.0.0/10 for Carrier-Grade NAT, separate from the private addresses of RFC 1918, because it is for the provider’s network. Such packets are not forwarded past the provider’s border. Geolocation locates the CGN server, not the end device. Several subscribers share one external public IPv4. In the console example, two subscribers with one external address cannot connect to each other. There is no count of subscribers per address. This is not a data-center label and not a “residential proxy”: the second phrase is not in the RFC.

CGN and an ASN are easy to mix up, because both are about “many behind one name”. An ASN is the number of a whole system’s routing policy. An external IPv4 under CGN is one translation address for several subscribers. A data center in the MRC examples is a known source of invalid activity, recognized by a list, not by a number from RFC 1930.

Networks Yandex names as its own

The Webmaster help says Yandex’s robots use the autonomous networks AS13238, AS208722, and AS212066. Addresses change often, so an IP list is not published. The help prefers a reverse-DNS check to an address list. The presence of a Yandex ASN does not mean every address in that system is a search robot.

Some robots download documents for something other than indexing and may not obey User-agent: *. YaDirectFetcher does not use the robots.txt file. Metrica’s robots are marked in the help table as not following the general robots.txt rules. The Metrica 4.0 variant downloads and caches CSS in order to reproduce the page in Webvisor. The records YandexAdditionalBot and YandexAdditional are honored in robots.txt so that content can be limited in quick answers with YandexGPT and in Search answers with Alice. They do not make requests for indexing.