- Efficient join for preference queries
- The project aims to implement an efficient join algorithm for preference queries (e.g., Skyline). We used a filter-refine approach, to early discard tuples from both relations that are not part of the join set. We have dramatically reduced the comparison compared to the state-of-the-art algorithm. Moreover, our algorithm gives earlier results. We have implemented the proposed algorithm in the PostgreSQL database.
- Location privacy techniques based on space transformation
- The project aims to explore space transformation as tool to maintain user privacy. This approach gives a higher level of security while presenting an efficient query processing. We transform object location and queries into a transformed space. We aim to find a reasonable space transformation method that balances between performance and privacy preserving.
- Integrating Database system with object-oriented storage devices
The project goal is to make the databases work efficiently with object-oriented storage devices (OSD). We investigate the OSD interface, and implement a prototype within MySQL database. The project is supported by DISC group, DTC, university of Minnesota.
- Skyline query processing for data with imprecise values
Given a large search space of multi-dimensional points, a skyline query finds points that represent a responsible choice. A point p is said to dominate point q if p is better than or equal q in all dimensions while p is strictly better than q in at least one dimension. All previous algorithms ignored that fact that real world data are imperfect and in many cases are incomplete. In our research, we develop the first skyline query processing algorithm that tolerates incomplete (i.e., missing) data and/or uncertain data (values may span a certain range). A working implementation for these ideas has been developed inside the PostgreSQL open-source database.
- Non blocking join operator
The objective of this project is to develop and implement an efficient non-blocking join operator inside the PostgreSQL open-source database management system. The main idea of being non-blocking is to have the ability to produce partial progressive query results even data sources are not completely read or have been blocked due to slow or bursty network connections. The operator based on probability model chose either to produce joined tuples form memory or from the disk. Thus, the user still receives progressive query results even incoming data are temporarily blocked. Experimental evidence shows that our developed non-blocking join algorithm performs order of magnitude better than state-of-the-art non-blocking join algorithms for the case of frequently blocked data sources.
- Hash function analysis, Master thesis, Alexandria University, Egypt
My master thesis includes two main contributions: (1) To protect SHA hash family from an adversary attack. The main idea is to add a wrapper for the SHA hash function that receives the input of the hash function and extends it in way that invalidates existing adversary attack. The output of the wrapper is considered as the new input to the SHA hash function. Our solution can be integrated in both software and hardware implementation. (2) To devise a new attack on MD2 hash function with the order O(297). Our attack outperforms the state-of-the-art attacks that have the order O(2104).
The views and opinions expressed in this page are strictly those of the page author.
The contents of this page have not been reviewed or approved by the University of Minnesota.