Authenticating Aggregate Queries over Set-Valued Data with Confidentiality
The main aim of the project to make a decision with clear manner in aggregate query over the dataset to search the overall dataset and find the correct result and show the client so decision process so easy manner.
We consider two potential security threats: 1) the SP could provide unfaithful query execution, thereby returning incorrect or incomplete query results; and 2) data privacy could be breached if sensitive source data are disclosed to the query client. Thus, the authentication problem we are investigating is for the query client to verify that the SP executes Q faithfully in terms of the following conditions: 1) the candidate objects are correctly selected and no objects in the selection range are skipped; 2) the returned features and multiplicities are not tampered with; and 3) the query result satisfies the aggregation semantics. The confidentiality requirement in this problem is to protect the objects’ (sensitive) feature sets against.
In this paper we study authenticated aggregate query services over set-valued data with confidentiality preservation. We assume that a dataset consists of one sensitive set-valued attribute (e.g., mutation-gene set) and multiple non-sensitive attributes (e.g., ZIP code and age). As illustrated in Example 1, an aggregate query is defined as a query whose result is derived from aggregates of data. In this paper, each aggregate query consists of two phases: a filtering phase that filters data by a range selection on non-sensitive attributes, followed by an aggregating phase that aggregates on the sensitive set-valued attribute. For broader applicability, we model the sensitive set-valued attribute as a multiset, i.e., a set that allows duplicate elements and to remove the duplicate value.