5th Year MS Thesis Presentation - Justin Zhang

— 12:00pm

Location:
In Person - Gates Hillman 7501

Speaker:
JUSTIN ZHANG, Master's Student, Computer Science Department, Carnegie Mellon University
https://www.justinz.dev/

Secure Convertible Codes

Large-scale distributed storage systems (DSS) make use of erasure codes to enforce fault tolerance in the event of node failure.  Due to observed changing failure rates within these systems, code redundancy tuning, or code conversion has been shown to reduce storage cost.  Prior work has developed bounds and constructions across many parameters for convertible codes, a class of erasure codes optimizing either the access or bandwidth costs of conversion. 

In this thesis, we investigate the information-theoretic security of convertible codes under the presence of an eavesdropper whom we enforce to learn nothing of the stored message. While current convertible code constructions are inherently insecure since they are systematic, we present novel constructions that augment existing cost-optimal convertible codes with perfect eavesdropper security. 

Furthermore, we prove that our constructions maximizes the amount of information that can be stored on such a system, and we give additional constructions and bounds for secure codes when additional information is known about the distribution of eavesdropped nodes. 

Thesis Committee

Rashmi Vinayak (Chair)
Aayush Jain

Additional Information