11. User Logs

11.1. Background information

Updates to DISQOVER are released regularly, and streamlining the user experience is a key aspect of improving the software. Information on how customers use DISQOVER, analyzed in an aggregated way, assists future product improvements.

In order to protect the privacy of the users and confidentiality of customer data, DISQOVER collects this information in a secure, controlled and anonymized way. Some information is not collected: specific search terms, names of individual users, names of local (customer-owned) canonical types, facets and properties. This data is anonimyzed in a non-reversible way on the customer’s server, and sent securely to storage at ONTOFORCE.

ONTOFORCE has committed itself to never attempt to reverse this hashing at any point, to never share this data with any third party at any point.

11.2. Technical implementation

Periodically, the customer’s local log files are digested according to the following steps:

1. Only lines with an [ACTION] tag are retained, as they indicate the interaction of the user with
DISQOVER.
2. The following pieces of information are salted with a key and hashed:
  • user email address
  • customer local facets (uri and label)
  • customer local canonical types (uri and label)
  • customer local relation types (uri and label)
  • customer local properties
  • customer local datasets
  • facet values filtered on (local and public ones)
  • instance ids and labels (local and public ones)
3. A full list of users is exported. For each user an anonymized (hashed) email address, the creation
date, the last login times and user rights is exported. For the current global configuration the number
of local canonical types, facets and properties are exported.
4. All these exports are tarred and zipped and are sent to ONTOFORCE storage.

11.3. Important notes

Canonical types, facets, and properties which contain the ONTOFORCE namespace ‘ns.ontoforce.com’ in their URI are considered public and non-confidential data, and are not hashed.

The hashing is performed using the MD5 checksum hash method. The salt key used by default is the server MAC address, but can be changed by the customer (see manual for further details). The salt key never leaves the customer’s server. Consequently, the following information from the log is available to ONTOFORCE:

  • Unique but anonymized identifiers (because of the hashing) for the following data:
    • User email address.
    • Names of customer local facets, canonical types, properties and datasets.
  • A date time stamp for each action.
  • The session id.
  • Which actions each unique user performs.
  • Names of public facets, canonical types, properties and datasets used by the user