Skip to content

URL access pattern detector #3

@krizhanovsky

Description

@krizhanovsky

Use case

A shopping bot cart/inventory hoarding & automated checkout , specific with the large number of requests to add to cart.

The bot may use thousands of IPs and different fingerprints, thanks to residential proxies.

The large number of cart requests means the large number of requests to cart within one HTTP session from the same IP and/or prevailing number of cart requests to over requests, such as goods browsing.

Trigger

Develop a trigger as a threshold for particular URLs. Many URLs and regexps should be allowed. All of URL pattern should be able to be specified with different thresholds per different time frames, such as minutes or seconds

Detector

For each of the client, identified by Tempesta Fingerprints or IP, compute the ratio of URL accesses from the trigger to all others and block the top clients.

Scoring

As a client traverses URLs, excluding the triggers, each of the request adds negative value to the score. The increment value should be derived from training data by requests preceding cart.

Metadata

Metadata

Assignees

No one assigned

    Labels

    Projects

    No projects

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions