Scalable Data Management Techniques and Parallel Programming Models for Key-Value Based Storage Systems for Cloud Computing

Abstract: 

Scalable Data Management Techniques and Parallel Programming Models for Key-Value Based Storage Systems for Cloud Computing

This project uses Itasca for two research projects. The goal of the first project is to investigate transactional data management primitives for key-value based storage systems, such as Hadoop/HBase, with the aim of supporting serializable transactions. The goal of this work is to develop scalable techniques for serializable transactions using the snapshot isolation model. The goal of the second project is to develop a scalable parallel programming framework based on key-value data storage models. The focus of this work is on graph problems that exhibit amorphous parallelism and which are not typically suitable for MapReduce based programming models. 

A bibliography of this group’s publications acknowledging MSI is attached.

Group name: 
tripathi
Attachment: