It still uses the old MS Indexing Service. However, it used to be done by getting a full list of matching documents from Indexing Service through ODBC, and a full portal document list, all into memory. Then the 2 lists would be matched. The problem was server resources with 10+k documents in a portal. With the overhaul, the Indexing Service data is extracted not by the module, but by SQL Server. This way, the whole search and match can be done through SQL (no need for ODBC any more). This is a linked server solution. Note that in some installations I've experienced the used SQL account needs db admin rights. I hope to have more documentation on this shortly.
Peter
PS The quick search box is another item. It used to be 'standard', then I removed it (there is a way to integrate DMX search with DNN search now), and now it is optional as some people wanted to keep it.
|