Database Reference
In-Depth Information
can apply security to segments of text within a document, including fields. A
given document, although indexed only once, can appear differently to dif-
ferent groups of users based on their security settings (think back to the busi-
ness plan example earlier in this chapter). Such security settings are not pos-
sible with vector space indexing, because users either have access to the
whole document or to none at all, limiting an organization's flexibility in
sharing information across the enterprise.
Index Auditing
For Big Data applications that require detailed auditing and accounting, the
ingestion of data into indexes can be fully audited through Data Explorer's
audit-log function. Data Explorer generates an audit-log entry for each piece
of content being sent for indexing. The log entry is guaranteed to contain all
of the errors and warnings that were encountered during crawling and
indexing. Using the audit log, Data Explorer deployments can ensure that
100 percent of the content is always accounted for: each item is either
indexed or has triggered an error that was reported to the administrator.
Such content-completeness audits are a key requirement for many legal and
compliance discovery applications.
User Management Layer
The User Management Layer includes all of the resources that users and
applications need in order to interact with data that's indexed by Data Ex-
plorer. Most importantly, this includes an interface to the Data Explorer
search engine, which handles all of the data requests from connected sources.
Before querying any data, users must authenticate. The user 's identity and
group affiliations for each repository that has been indexed are stored in
user profiles or accessed at login time from a directory service. (We'll note
that if you are using LDAP or Active Directory services, this information is
automatically retrieved.) After they are logged in, users can have a person-
alized interface that reflects their profile.
This layer also contains Data Explorer's ability to federate queries to exter-
nal sources that are not natively indexed by Data Explorer, such as premium
subscription-based information services on the Internet. The results can be
merged with native Data Explorer results to create an enriched and expanded
view of relevant information.
