Projects

From NetSysLab

Jump to: navigation, search


Main Current Projects

Graph processing

Dynamic Graphs -- Support for processing time-evolving graphs (collaboration with CASC@LLNL)
Pattern Matching in large metadata graphs (collaboration with CASC@LLNL)
Totem -- Graph processing on heterogeneous CPU and GPU platforms

Fault tolerant HPC applications

Error resilience -- Reliability: characterization and fault tolerance mechanisms for parallel applications
LetGo -- Forward recovery for HPC applications (in collaboration with LANL)

Cybersecurity/Social Networks

Artemis -- Defending against large-scale cyber intrusions by focusing on the vulnerable population (collaboration with LERSEE, Telus, Google)
Socialbots and Social Networks -- Cyber attacks in online social networks (collaboration with LERSEE)
Cheating, Online games, and Social Networks -- Cheating in the Stream community (collaboration with DSG@USF)

Harnessing accelerators (GPUs)

MUMmerGPU++ -- Genome sequence alignment on GPUs
Totem -- Graph processing on heterogeneous CPU and GPU platforms
StoreGPU -- Exploiting graphics processing units (GPUs) to accelerate distributed storage
InSAR Processing -- Accelerating kernels in the InSAR processing chain using multi-core CPUs and GPUs (collaboration with 3vGeomatics)
Harnessing integrated CPU/GPU platforms -- Exploring Nvidia Jetson.


Distributed Storage

MosaStore -- Storage system integrating contributory storage
ExM -- System support for extreme-scale, many-task applications (collaboration with CI@UChicago and ANL)
VMflock -- Virtual machine migration across clouds (collaboration with Storage Systems @IBM Almaden])
StoreGPU -- Exploiting graphics processing units (GPUs) to accelerate distributed storage


Past Projects

Social Tagging -- Characterizing individual and social behavior in online peer production systems
Crowdsourcing-based Applications --Exploring the basic design principles applicable to an array of crowdsourcing-based applications by investigating a specific case study, namely Smart Parking
ThriftStore -- Exploring data reliability tradeoffs in replicated storage systems
Data Distribution Strategies for Large-scale Scientific Data Collaborations
High-performance authorization infrastructures
CrystalGPU -- Transparent and efficient utilization of GPU power