Skip Ribbon Commands
Skip to main content
 
Research
 
The facility is being utilized in computation-intensive research projects in the areas of Fluid Dynamics and Biosciences, huge data processing applications such as Flood and Weather Forecasting, Financial Analysis, Oil and Gas Exploration, Energy Efficient Building Designs and Transportation Management at national-level. Some of the ongoing research projects are given below.

1 - Parallel Model Checking Tools and Techniques

Model-Checking technique developed by Clarke and Emerson is used for automated verification of finite state reactive systems. The specification of the system being investigated is expressed in temporal logic formula which is verified by exploring entire state space of the system. Model checking has prevailed as preferred methodology during the last three decades for verification of digital systems, communication protocols and biological regulatory networks. With rapid progress in Information and communication technology, the size and complexity of digital systems has increased exponentially and their transition graph exceed beyond billions of states. Complex biochemical processes, for instance involve millions of cells interacting with each other.
Current Model Checking Algorithms do not scale with increasing complexity of these systems and suffer from serious performance and memory constraints. The aim of this work is to develop parallel algorithms for Modern High Performance Computing platforms which exhibit great heterogeneity in terms of computational speed, memory hierarchies, size and organization.

2 - Modeling and Analysis of Gene Regulatory Networks on General Purpose Graphics Processors

Gene Regulatory Networks (GRN’s) are used to understand the functionality of organisms on micro level. Genetic regulatory networks of interest involve numerous components connected through many interlocking positive and negative feedback loops resulting in compute intensive state space. The approaches used to model these GRN’s use ordinary and partial differential equations, qualitative differential equations, artificial neural networks, Boolean networks, Bayesian networks, stochastic equations and rule-based formalisms. All of these approaches use huge data sets and complex algorithms for modeling and analysis of GRNs and suffer from performance drawbacks. The use of General Purpose Graphical Processing Units (GPGPUs) for modeling and analysis of GRNs seems very promising for modeling and analysis of GRN’s due to their low cost, suitability for data parallel problems and ease of programming. GPUs are already being used in computational biology especially in the field of protein sequence analysis. The aim of this work is to implement GRN analysis algorithms on GPUs.

3 - Performance Enhancement of Signature based Intrusion Detection Systems

Signature-Based Intrusion detection systems operates by comparing packet payloads against attack signatures. The process of signature matching takes up a lot of processing time and thus overwhelms the efficiency of a serial Intrusion Detection system due to exponential increase in number of vulnerabilities, network traffic and bandwidth. The motivation behind this work is to explore and develop different techniques for enhancing the performance of signature-based IDS.

4 - Development of Bulk Encryption Framework for Modern HPC Platforms