Many data models especially for those machine-generated data can be naturally represented as multidimensional arrays. For example, the four axes of climate data can correspond to a geographic location, altitude and time. Array database management system (ADBMS) is a new kind of DBMS which is designed to store and manipulate those kinds of data. On the other hand, information security is more and more important, and multidimensional data can be very sensitive, e.g. personal photos stored as 2-dimensional arrays on public servers. However, with the ubiquity of multidimensional arrays and cloud storage, the security has not gained enough attention. One possible way is to naively encrypt local array database before sending it to public servers. However, that will lose the accessibility to a particular part of the data.
Searchable Encryption (SE) is a technology that enables a data owner to outsource a database to a server in a private manner, so that the latter can still answer some kinds of queries without learning too much information about the underlying database and the queries.
Therefore, combining SE and array databases may be an ideal solution to ensure the confidentiality.
In this project, students are first to survey the basic principles of array databases, then understand the theory of SE and implement SE on existing ADBMSs, and finally optimize the design of previous SE according to the features of multidimensional arrays. For UROP students participating in this project, more practical tasks will be distributed because they can enhance their programming ability learnt from courses and understand the theories behind the project more concretely.
In this project, you will:
Improve your programming skills with full instructions and powerful machines.
Get the experience of installing and utilizing several different cutting-edge systems.
Opportunities to solve more challenging theoretical problems with our research group.
Excellent performance would lead to potential publications in quality conferences/journals.