ASN (Autonomous System Number)

An ASN is the identifier assigned to an autonomous system, meaning a network operator that announces its IP ranges to the rest of the internet over BGP. Amazon, a regional cable provider and a Frankfurt bare-metal host each hold their own. Every IP address a scraper uses belongs to exactly one autonomous system at any moment, and that mapping is public: anyone can look up the ASN behind an address in seconds.

Anti-bot systems use it as a first-pass classifier because it is cheap and it correlates well. Traffic from a known cloud ASN gets scored as automation by default, since almost nobody browses a retail site from an EC2 instance or a Hetzner box. Consumer ISP ASNs get no such penalty. In between sit hosting networks that historically leased space to proxy resellers, and those penalties are sticky: reputation attaches to the announcing network, so a range can stay suspect long after the abusive tenant left. This is why a datacenter pool can fail a target instantly while a residential pool of the same size walks through, with no difference in the scraper itself.

ASN targeting is also a purchasing feature. Bright Data's documentation shows it as a username parameter, `-asn-<number>` appended to the zone string, and states it is supported on residential proxies only (read 2026-07-27). SOAX advertises targeting by country, region, city or ASN on its residential product (read 2026-07-27). The useful part is not the routing control, which most teams never need. It is the diligence question the feature implies: ask a vendor for the ASN distribution of its pool, not just the country breakdown. Country coverage is easy to inflate. ASN spread is hard to fake because you can verify it yourself from sampled exits.

Read that spread as the honest measure of pool quality. A pool that resolves to a long tail of consumer ISPs across many operators behaves like consumer traffic under pressure. A pool that concentrates in three hosting ASNs will fail on defended targets no matter how many million IPs the marketing page claims, because the defender is scoring the network, not the address. Collect exit IPs over a trial window, resolve each to its ASN, and look at what share of your sample the top three account for. If the answer is most of it, the headline number is describing volume within a handful of neighborhoods.