7.6 Federated Searches
Search activity generated by federated search engines MUST be categorized separately from searches conducted by users on the host platform.
Any searches generated from a federated search system MUST be included in the separate Searches_Federated counts within Database Reports and MUST NOT be included in the Searches_Regular or Searches_Automated counts.
The most common ways to recognize federated search activity are as follows:
A federated search engine may be using its own dedicated IP address, which can be identified and used to separate out the activity.
If the standard HTML interface is being used (e.g. for screen scraping), the user agent within the web log files can be used to identify the activity as coming from a federated search.
For Z39.50 activity, authentication is usually through a username/password combination. Create a unique username/password that just the federated search engine will use.
If an API or XML gateway is available, set up an instance of the gateway that is for the exclusive use of federated search tools. It is RECOMMENDED that you also require the federated search to include an identifying parameter when making requests to the gateway.
COUNTER provides lists of user agents that represent the most common federated search tools. See Appendix G.