Modern life is quickly becoming inundated with a flood of information and data feeds from all sources. Traffic networks, healthcare systems, wireless networks, electrical power grids, social media networks—the information from any one of these alone is staggering, but what happens when all the networks must be considered together? How are we to sift through the deluge to find the most useful data and to understand how these massive systems interact?
One of the most important needs to arise from network science is the need for middleware, a user interface layer that helps organize and interpret the findings of complex calculations. Because systems are so massive and their interactions are so complex, high-performance computing capabilities are necessary to deal with the volume and variety of information.
CyberInfrastructure for NETwork science, or CINET, provides the vital infrastructure needed for network science with a suite of applications that helps users analyze network structure and dynamics. It also provides an environment for collaborative science and acts as a community resource. Community members can contribute course materials, networks, analyses, algorithms, simulations and other resources.
A social network powered by CINET
“Network science is the study of networks,” says Keith Bisset, Senior Research Scientist, “which seems simple but is fairly complex and pervasive.” It’s not just about understanding the shape of the network or how it interacts with other networks, but which nodes are the most information-rich in any given network. For instance, when studying network-based epidemiology, it’s important to understand which nodes have the greatest influence in the system and how different methods for dealing with those nodes might alter the system.
CINET was designed to be very user-friendly. Users of CINET benefit from an open access system where they can customize networks and apply a variety of models and algorithms to extend their research capabilities. The CINET system offers a simple computational interface for a variety of domain experts and houses incredibly large networks due to its data management facilities and flexibility.
Not only has it been used in network science courses, but it is also accessible to a variety of specialists from various disciplines. CINET’s layered architecture provides several key tools for users: Granite for network structural analysis, GDSC (Graph Dynamical Systems Calculator) for phase space analysis of graph dynamics, and EDISON for network dynamics and the spread of contagion over networks.
CINET Workshop Video, July 2015.
“Initially, CINET consisted of a graph library, but has subsequently evolved to incorporate several graph libraries from other institutions, such as the SNAP library from Stanford and the NetworkX library from Los Alamos,” says Chris Kuhlman, Research Scientist at the Network Dynamics and Simulation Science Laboratory. “Future plans call for incorporating still more libraries, so in this sense, the goal is to devise and develop a community resource where many tools are provided to the user and they can select from the tools that can best accomplish their given tasks.”
CINET is the result of collaborations with a variety of institutions and renowned laboratories such as Indiana University, North Carolina Agricultural and Technical State University, University at Albany State University of New York and University of Houston-Downtown, among others. Teams of scientists have built CINET’s scalable architecture as a means of opening up the field of network science to researchers and students who might not otherwise realize its benefits to their fields.
The real-world applications of CINET are almost endless; cybersecurity, fraud detection, and epidemiology are just a few areas that have benefitted from this revolutionary cyberinfrastructure tool. Soon CINET will have the necessary enhancements that will allow researchers to study more dynamic models, “contagions” as they’re called, such as how information spreads through a social network like Twitter or through the neural network of the brain.
In this way, we can “kayak the firehose of information,” as Research Assistant Professor Allan Dickerman says, rather then being drowned by it.
Workshop participants at the CINET workshop in Blacksburg, July 10, 2015.
On July 10, a workshop for CINET users was held in Blacksburg. Participants and members learned more about how CINET is used in applications ranging from cybersecurity to education. Another workshop will take place at University at Albany State University of New York on August 11. Register here!
As the utility of CINET continues to grow, there are plans for additional workshops in the future.