The I/UCRC Center for Cloud and Autonomic Computing (CAC) developed a comprehensive research and education program, in close collaboration with industry and government members, to address issues of design, use and management of cloud computing, IT systems and IT application complexity through autonomic approaches. Autonomic computing approaches enable systems and applications to manage themselves, making them more reliable, more secure, and more efficient. The center conducted scientific and engineering research and development on methods, architectures and technologies for the design, implementation, integration and evaluation of special- and general-purpose computing systems, components and applications that are provisioned by IT clouds and/or are capable of autonomously achieving desired behaviors. The center sites, distributed across four institutions (University of Florida, University of Arizona, Rutgers University, and Mississippi State University), also educated students in the interdisciplinary fields of cloud and autonomic computing. Specifically, the center addressed research in the domains of cybersecurity, management of virtualized systems (virtual machines, virtual storage, and virtual networks), energy conumption in data centers, and high-performance distributed computing applications. Outcomes of the research activities included papers in journals and conferences, technology transfer to inudstry, and open-source software. Highlited outcomes of center projects include the following. A startup company, AVIRTEK, was founded to transition the University of Arizona autonomic management and cybersecurity solutions to the marketplace, and cybersecurity technologies developed at the UA CAC site have transitions to this company. The company has started deploying commercially network appliances with automated and integrated management for small and medium size networks. Power conumption is an increasingly significant percentage of the cost of operating large data centers, which are used by banks, investment firms, IT service providers and other large enterprises. A project at the University of Florida researched virtualization-based autonomic computing approaches to monitor, model and predict workloads associated with individual servers; model and predict global resource demand; dynmically allocate and de-allocate virtual machines to physical machines; device methods based on control theory and/or market-based approaches to use the above-described  mechanisms to minimize the cost of providing individual services while globally minimizing power consumption and delivering contracted service levels. Research also considered self-organizing virtual networks and efforts to connect personal devices of a user and acrosss their social network, leading to development of an open-source infrastructure that allows users to create Virtual Private Networks (VPNs) connecting resources from different sites securely, without the cost and complexity associated with setup and management of typical approaches.

Contributions of this infrastructure include techniques associated with self-configuraiton of virtual networks ofr the the simple deploymeht of collaborative environments.  The research projects at the Rutgers University site have demonstrated how autonomic computing can effectively optimize various aspects of computing systems and applications, including efficiency, performance, adherence to Service Levels Agreements (SLAs), fault tolerance and cost-effectiveness. Also, Rutgers explored ways to take actions to reduce energy consumption at the server side in a large machine room before performing costly migrations of virtual machines (VMs). Specifically, Rutgers focussed on exploiting VM-based configurations which are complementary to other techniques at the physical server layer, such as low-power modes. Results obtained from an experimental evaluation on ...

Please report errors in award information by writing to: awardsearch@nsf.gov.