Phase II IUCRC CRSS: Center for Research in Storage Systems

University Of California-Santa Cruz, Santa Cruz CA

Investigators

Abstract

The Center for Research in Storage Systems (CRSS) conducts research to improve the performance, security, reliability, efficiency, and usability of data storage. A system composed of thousands or even millions of storage devices must provide these characteristics, though individual devices may lack them. Finding and using data in such large, complex systems is challenging, yet it must be made simple and efficient for end users. Complex systems are also subject both to hardware and software failures as well as directed attacks to steal or destroy data, so storage systems must guard against both security breaches and data loss. CRSS is concerned with adapting systems to better utilize new storage technologies such as byte-addressable non-volatile memory (BNVM), shingled disk, and glass. Operating systems and file systems must adapt to these new technologies to provide higher performance and better usability for BNVM, and improved security, efficiency, and long-term reliability for archival storage. Cross-cutting issues such as improved security and the usage of machine learning to optimize storage systems are relevant for all types of storage systems, yet there remain many open questions in how to address these issues, particularly given the large scale in both time and size that CRSS is exploring. Improving storage systems for large-scale data science is also a significant challenge that CRSS is addressing. Data storage is the foundation of our digital society: the vast majority of our knowledge and information is stored digitally. The research done at CRSS will allow society as a whole to better keep and utilize the flood of information our world is gathering. CRSS research will help cloud providers provide more reliable and usable large-scale storage, and will help reduce the risk of security breaches. CRSS research will also help organizations store and access the massive amounts of data generated by the Internet of Things, further aiding a wide range of societal issues such as climate, transportation, and housing. CRSS will maintain a Web-accessible project repository at https://www.crss.ucsc.edu/repository.html, which will have links to publications, data, and code developed by all CRSS sites. Code will be hosted on either gitlab.soe.ucsc.edu or bitbucket.org. Traces will be hosted locally or donated to the repository operated by the Storage Networking Industry Association's Input/Output Traces, Tools, and Analysis (IOTTA) Technical Work Group. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.