Online Indexing for Databases Using Query Workloads
Source: Kurukshetra University
A query response in DBMS that is developed based on sequential search or static snapshot (or static indexing) may significantly degrade if the query patterns change and also with the increase in the database size. Online indexing addresses these issues by considering a parameterizable technique to recommend the indexes based on index types that are frequently used for data sets and to dynamically adjust indexes as the query workload changes. The two parameters the authors have considered are support (prediction of sale of frequent products in the future considering transactions of past) and confidence (the probability with which a product moves with respect to another product/ products in the frequent item set).