Google has revamped the hardware and some aspects of the software for the Google Search Appliance, which can now support 10 million files in a single box, up from 3 million in the previous version.
Google has also added new biasing features which allow corporate administrators to weight search results differently for different end users – for example, the marketing and engineering departments might get different documents returned first for the same keyword. This is termed front-end biasing.
Metadata biasing is also new with this release. That lets admins rank metadata such as create date or author on a sliding scale of importance, so that, for example, documents written by the CEO are returned first.
Data growth is making indexing and search necessary, according to Forrester Research analyst Stephanie Balaouras. “I would say, once you can measure your storage in hundreds of TBs,” data indexing will probably be necessary, she said.
“Below 100 TBs, you should have a rough understanding of how that data is broken down between email, unstructured files, and database data,” she added. “And with some basic policies, the most important information should be ingested into a content management system.”