The LDV approach to polyinstantiation

In a multilevel DBMS, there is a need for tuple polyinstantiation. It is conceivable that users at different security levels will have different, legitimate understandings of the current state of the enterprise modeled by the database. As long as the data in each tuple in the database is classified at the same level, the tuple polyinstantiation approach is adequate. However, there are times when it is reasonable for a tuple to be distributed across security levels because it contains individual data elements that are classified at different levels and the responsibility for maintaining the various portions of the tuple should properly be allocated to users operating at different security levels. The LDV approach is to use tuple polyinstantiation augmented with the ability to derive tuple values from lower level tuples to provide a method of handling multilevel tuples that does not require element polyinstantiation.>