Return to Research

Workplan 2018

Distributed Infrastructure Support for Workflow and Data Management

Design of Tools for Dataset Integration (Task 1): Next-generation scientific discoveries are at
the boundaries of datasets, e.g., across multiple science disciplines, institutions and spatial and temporal
scales. In the context of Deduce and the DALHIS collaboration, we will explore three key research
topics. First, our current work has explored domain-specific metrics and our goal in the next year will be
to evaluate the wider applicability and user perception of the data change metrics. In the context of the
Deduce framework, we will explore real-time data change evaluation and associated algorithms. Finally,
we will explore how data change knowledge can be used to make system decisions such as selective file
transfer. The complementary skills from the Myraids team on distributed systems and data-streaming
makes it an ideal topic to explore in the DALHIS collaboration.

Security in HPC and Cloud Environments (Task 2) : We will jointly continue the work on data integrity in HPC systems that we  started during Amir Teshome Wonjiga’s first internship (2017).  In a second internship (2018), we plan to implement a prototype of the proposed system and perform an experimental evaluation. Regarding the work on building a data analysis workflow for anomaly detection, we plan to improve our current model and evaluate it.  We also plan to examine ways in which advanced encryption techniques and privacy-preseving technologies — such as fully homomorphic encryption, secure multiparty computation, functional encryption, and differential privacy — can enhance both the security and usability in which science is conducted.