Recoverability and Reliability

Recoverability and reliability are critical. If a system has a hardware failure and goes down in the middle of indexing five million documents, it isn't efficient to be forced to begin re-indexing the entire corpus. So great effort was put into making sure that the indexing system is recoverable. When a system comes up after a failure and disk integrity is verified, Index Server can then be activated to do a full integrity check of all indexes. Using the same kind of logging and recoverability mechanism that the file system uses, Index Server can detect corruptions and fix them.

Index Server works in a transaction-oriented manner when writing data to the disk, which avoids indexing only half a document. Either the entire document is there, and guaranteed to be correct once the data is indexed, or the document isn't accepted. This eliminates any inconsistent states due to partial indexing.

Index Server is built upon the solid and robust foundation of the Windows NT operating system to assure reliable service in a full 24-hour by 7-day operation on a mission-critical server.