ECE 6102: Dependable Distributed Systems
Some Past Projects
You are encouraged to develop your own project ideas and discuss them with me.
The constraints are that it must involve distributed computing, it must entail
substantial C/C++ and/or Java programming (no primarily Perl, Tcl, or Visual-Basic
projects accepted), and it must have a significant part dealing with security,
fault tolerance, or general robustness. Primarily Python projects will not be
specifically disallowed but will be carefully scrutinized. Cloud-based
projects are strongly encouraged.
-
Reliable Stream Analysis on the Internet of Things
Live stream video processing for resource-poor devices. The final
report for this project is included as an example of an "A" project.
IOT Stream Analysis Final Report
-
Dependable Storage: Fault-tolerant Secure Distributed File System
Block-oriented
distributed storage with replication and encryption.
-
GeoShare: Geographically-Distributed and Secret Shared Data Storage in
the Cloud
Secure cloud storage built on top of Amazon S3 with secret shares
distributed across Amazon regions world-wide.
-
Virtualization: Support for Software Replication
Modification of a virtual machine monitor to provide totally ordered
message delivery, deterministic execution, etc., in order to provide
support for software replication. Demonstration of a simple replicated
process execution on the chosen platform.
-
Dependable Storage: Secure FTP Repository
Design and implementation of a distributed FTP repository using proactive
secret sharing techniques to maintain confidentiality and integrity. Includes
an approach to deal with nearly-coincident attacks on all repository
sites.
-
Mobile Devices: File System Verification and Repair on
Resource-Constrained Devices
Maintain distributed integrity information about a mobile device file
system, detect unauthorized changes to file system, and repair corrupted
files with help of other devices and/or servers.
-
Distributed Whois Crawler
A distributed crawler that gathers data through domain information groper
(dig) requests to whois servers. The project builds upon a Chord
architecture to provide redundant archives of the gathered data across
multiple resources. A convenient, RESTful interface allows querying of the
gathered data and coordinates with Chord to gather archived data.
-
A Dynamic Replica Management System for HDFS
Dynamically change the number of replicas of a file in the Hadoop
Distributed File System, based on the usage and/or importance of each
file.